Approval Package For

Home , Regorafenib

CENTER FOR DRUG EVALUATION AND RESEARCH

Approval Package for:

APPLICATION NUMBER:

203085Orig1s007

Trade Name: Stivarga

Generic or Proper regorafenib Name:

Sponsor: Bayer HealthCare Pharmaceuticals, Inc.

Approval Date: April 27, 2017

Indication: STIVARGA is a kinase inhibitor indicated for the treatment of patients with: • Metastatic colorectal cancer (CRC) who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy, and if RAS wild-type, an anti- EGFR therapy. • Locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate. • Hepatocellular carcinoma (HCC) who have been previously treated with sorafenib.

CENTER FOR DRUG EVALUATION AND RESEARCH 203085Orig1s007

CONTENTS

Reviews / Information Included in this NDA Review.

Approval Letter X Other Action Letters Labeling X REMS Summary Review X Officer/Employee List X Office Director Memo Cross Discipline Team Leader Review Medical Review(s) X Chemistry Review(s) X Environmental Assessment Pharmacology Review(s) Statistical Review(s) X Microbiology / Virology Review(s) Clinical Pharmacology/Biopharmaceutics Review(s) X Other Reviews X Risk Assessment and Risk Mitigation Review(s) Proprietary Name Review(s) Administrative/Correspondence Document(s) X

CENTER FOR DRUG EVALUATION AND RESEARCH

APPLICATION NUMBER:

203085Orig1s007

APPROVAL LETTER

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Food and Drug Administration Silver Spring MD 20993

NDA 203085/S-007 SUPPLEMENT APPROVAL

Bayer HealthCare Pharmaceuticals, Inc. Attention: Lisa Chao, Ph. D. Deputy Director, Global Regulatory Affairs, Specialty Medicine 100 Bayer Boulevard, P.O. Box 915 Whippany, NJ 07981-0915

Dear Dr. Chao:

Please refer to your Supplemental New Drug Application (sNDA) dated October 30, 2016, received October 31, 2016, and your amendments, submitted under section 505(b) of the Federal Food, Drug, and Cosmetic Act (FDCA) for Stivarga (regorafenib) tablets, 40 mg.

This Prior Approval supplemental new drug application provides for a new indication for the use of Stivarga (regorafenib) for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with sorafenib.

APPROVAL & LABELING

We have completed our review of this supplemental application, as amended. It is approved, effective on the date of this letter, for use as recommended in the enclosed, agreed-upon labeling text.

WAIVER OF HIGHLIGHTS SECTION

We are waiving the requirements of 21 CFR 201.57(d)(8) regarding the length of Highlights of prescribing information. This waiver applies to all future supplements containing revised labeling unless we notify you otherwise.

CONTENT OF LABELING

As soon as possible, but no later than 14 days from the date of this letter, submit the content of labeling [21 CFR 314.50(l)] in structured product labeling (SPL) format using the FDA automated drug registration and listing system (eLIST), as described at http://www.fda.gov/ForIndustry/DataStandards/StructuredProductLabeling/default.htm. Content of labeling must be identical to the enclosed labeling (text for the package insert, text for the patient package insert), with the addition of any labeling changes in pending “Changes Being

Reference ID: 4091525 NDA 203085/S-007 Page 2

Effected” (CBE) supplements, as well as annual reportable changes not included in the enclosed labeling.

Information on submitting SPL files using eList may be found in the guidance for industry titled “SPL Standard for Content of Labeling Technical Qs and As at http://www.fda.gov/downloads/DrugsGuidanceComplianceRegulatoryInformation/Guidances/U CM072392.pdf

The SPL will be accessible from publicly available labeling repositories.

Also within 14 days, amend all pending supplemental applications that include labeling changes for this NDA, including CBE supplements for which FDA has not yet issued an action letter, with the content of labeling [21 CFR 314.50(l)(1)(i)] in MS Word format, that includes the changes approved in this supplemental application, as well as annual reportable changes and annotate each change. To facilitate review of your submission, provide a highlighted or marked- up copy that shows all changes, as well as a clean Microsoft Word version. The marked-up copy should provide appropriate annotations, including supplement number(s) and annual report date(s).

REQUIRED PEDIATRIC ASSESSMENTS

Under the Pediatric Research Equity Act (PREA) (21 U.S.C. 355c), all applications for new active ingredients, new indications, new dosage forms, new dosing regimens, or new routes of administration are required to contain an assessment of the safety and effectiveness of the product for the claimed indication(s) in pediatric patients unless this requirement is waived, deferred, or inapplicable.

Because this drug product for this indication has an orphan drug designation, you are exempt from this requirement.

PROMOTIONAL MATERIALS

You may request advisory comments on proposed introductory advertising and promotional labeling. To do so, submit the following, in triplicate, (1) a cover letter requesting advisory comments, (2) the proposed materials in draft or mock-up form with annotated references, and (3) the package insert(s) to:

OPDP Regulatory Project Manager Food and Drug Administration Center for Drug Evaluation and Research Office of Prescription Drug Promotion (OPDP) 5901-B Ammendale Road Beltsville, MD 20705-1266

Reference ID: 4091525 NDA 203085/S-007 Page 3

Alternatively, you may submit a request for advisory comments electronically in eCTD format. For more information about submitting promotional materials in eCTD format, see the draft Guidance for Industry (available at: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/U CM443702.pdf ).

You must submit final promotional materials and package insert(s), accompanied by a Form FDA 2253, at the time of initial dissemination or publication [21 CFR 314.81(b)(3)(i)]. Form FDA 2253 is available at http://www.fda.gov/downloads/AboutFDA/ReportsManualsForms/Forms/UCM083570.pdf.

Information and Instructions for completing the form can be found at http://www.fda.gov/downloads/AboutFDA/ReportsManualsForms/Forms/UCM375154.pdf.

For more information about submission of promotional materials to the Office of Prescription Drug Promotion (OPDP), see http://www.fda.gov/AboutFDA/CentersOffices/CDER/ucm090142.htm.

REPORTING REQUIREMENTS

We remind you that you must comply with reporting requirements for an approved NDA (21 CFR 314.80 and 314.81).

If you have any questions, call Anuja Patel, Senior Regulatory Health Project Manager, at (301) 796-9022.

Sincerely,

{See appended electronic signature page}

Patricia Keegan, M.D. Division Director Division of Oncology Products 2 Office of Hematology and Oncology Products Center for Drug Evaluation and Research

ENCLOSURE: Content of Labeling

Reference ID: 4091525 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------PATRICIA KEEGAN 04/27/2017

Reference ID: 4091525

CENTER FOR DRUG EVALUATION AND RESEARCH

APPLICATION NUMBER:

203085Orig1s007

LABELING HIGHLIGHTS OF PRESCRIBING INFORMATION  Infections: Withhold STIVARGA in patients with worsening or severe These highlights do not include all the information needed to use infections. (5.2) STIVARGA safely and effectively. See full prescribing information for  Hemorrhage: Permanently discontinue STIVARGA for severe or life- STIVARGA. threatening hemorrhage. (5.3) STIVARGA® (regorafenib) tablets, for oral use  Gastrointestinal perforation or fistula: Discontinue STIVARGA. (5.4) Initial U.S. Approval: 2012  Dermatologic toxicity: Withhold and then reduce or discontinue

STIVARGA depending on severity and persistence of dermatologic WARNING: HEPATOTOXICITY toxicity. (5.5) See full prescribing information for complete boxed warning.  Severe and sometimes fatal hepatotoxicity has occurred in clinical  Hypertension: Temporarily or permanently withhold STIVARGA for severe trials. (5.1) or uncontrolled hypertension. (5.6)  Monitor hepatic function prior to and during treatment. (5.1)  Cardiac ischemia and infarction: Withhold STIVARGA for new or acute  Interrupt and then reduce or discontinue STIVARGA for cardiac ischemia/infarction and resume only after resolution of acute hepatotoxicity as manifested by elevated liver function tests or ischemic events. (5.7) hepatocellular necrosis, depending upon severity and persistence.  Reversible posterior leukoencephalopathy syndrome (RPLS): Discontinue (2.2) STIVARGA. (5.8)  Wound healing complications: Discontinue STIVARGA before surgery. ------RECENT MAJOR CHANGES ------Discontinue in patients with wound dehiscence. (5.9) Indications and Usage, Colorectal Cancer (1.1) 6/2016 Indications and Usage, Hepatocellular Carcinoma (1.3) 4/2017  Embryo-fetal toxicity: Can cause fetal harm. Advise women of potential Dosage and Administration, Dose Modifications (2.2) 4/2017 risk to a fetus and to use effective contraception during treatment and for 2 Warnings and Precautions (5.1-5.8) 4/2017 months after the final dose. Advise males to use effective contraception for 2 months after the final dose. (5.10, 8.1, 8.3) ------INDICATIONS AND USAGE ------ADVERSE REACTIONS ------STIVARGA is a kinase inhibitor indicated for the treatment of patients with: The most common adverse reactions (≥20%) are pain (including  Metastatic colorectal cancer (CRC) who have been previously treated with gastrointestinal and abdominal pain), HFSR, asthenia/fatigue, diarrhea, fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti- decreased appetite/food intake, hypertension, infection, dysphonia, VEGF therapy, and, if RAS wild-type, an anti-EGFR therapy. (1.1) hyperbilirubinemia, fever, mucositis, weight loss, rash, and nausea. (6)  Locally advanced, unresectable or metastatic gastrointestinal stromal tumor To report SUSPECTED ADVERSE REACTIONS, contact Bayer (GIST) who have been previously treated with imatinib mesylate and HealthCare Pharmaceuticals Inc. at 1-888-842-2937 or FDA at 1-800 sunitinib malate. (1.2) FDA-1088 or www.fda.gov/medwatch  Hepatocellular carcinoma (HCC) who have been previously treated with ------DRUG INTERACTIONS ------sorafenib (1.3)  Strong CYP3A4 inducers: Avoid strong CYP3A4 inducers. (7.1)  Strong CYP3A4 inhibitors: Avoid strong CYP3A4 inhibitors. (7.2) ------DOSAGE AND ADMINISTRATION ------ BCRP substrates: Monitor patients closely for symptoms of increased  Recommended dose: 160 mg orally, once daily for the first 21 days of each exposure to BCRP substrates. (7.3) 28-day cycle. (2.1) ------USE IN SPECIFIC POPULATIONS ------ Take STIVARGA after a low-fat meal. (2.1, 12.3) Nursing Mothers: Discontinue drug or nursing, taking into consideration the ------DOSAGE FORMS AND STRENGTHS ------importance of the drug to the mother. (8.3) Tablets: 40 mg (3) See 17 for PATIENT COUNSELING INFORMATION and FDA- approved patient labeling. ------CONTRAINDICATIONS ------None. Revised: 4/2017 ------WARNINGS AND PRECAUTIONS ------ Hepatotoxicity: Monitor liver function tests. Withhold and then reduce or discontinue STIVARGA based on severity and duration. (5.1)

FULL PRESCRIBING INFORMATION: CONTENTS* 6.2 Postmarketing Experience WARNING: HEPATOTOXICITY 7 DRUG INTERACTIONS 1 INDICATIONS AND USAGE 7.1 Effect of Strong CYP3A4 Inducers on Regorafenib 1.1 Colorectal Cancer 7.2 Effect of Strong CYP3A4 Inhibitors on Regorafenib 1.2 Gastrointestinal Stromal Tumors 7.3 Effect of Regorafenib on Breast Cancer Resistance Protein (BCRP) 1.3 Hepatocellular Carcinoma Substrates 2 DOSAGE AND ADMINISTRATION 8 USE IN SPECIFIC POPULATIONS 2.1 Recommended Dose 8.1 Pregnancy 2.2 Dose Modifications 8.2 Lactation 3 DOSAGE FORMS AND STRENGTHS 8.3 Females and Males of Reproductive Potential 4 CONTRAINDICATIONS 8.4 Pediatric Use 5 WARNINGS AND PRECAUTIONS 8.5 Geriatric Use 5.1 Hepatotoxicity 8.6 Hepatic Impairment 5.2 Infections 8.7 Renal Impairment 5 3 Hemorrhage 8.8 Race 5.4 Gastrointestinal Perforation or Fistula 10 OVERDOSAGE 5.5 Dermatologic Toxicity 11 DESCRIPTION 5.6 Hypertension 12 CLINICAL PHARMACOLOGY 5.7 Cardiac Ischemia and Infarction 12.1 Mechanism of Action 5.8 Reversible Posterior Leukoencephalopathy Syndrome 12 2 Pharmacodynamics 5.9 Wound Healing Complications 12 3 Pharmacokinetics 5.10 Embryo-Fetal Toxicity 13 NONCLINICAL TOXICOLOGY 6 ADVERSE REACTIONS 13.1 Carcinogenesis, Mutagenesis, Impairment of Fertility 6.1 Clinical Trials Experience 13 2 Animal Toxicology and/or Pharmacology 1

Reference ID: 4090114

14 CLINICAL STUDIES 16 HOW SUPPLIED/STORAGE AND HANDLING 14.1 Colorectal Cancer 17 PATIENT COUNSELING INFORMATION 14.2 Gastrointestinal Stromal Tumors 14.3 Hepatocellular Carcinoma (HCC) *Sections or subsections omitted from the full prescribing information are not listed.

Reference ID: 4090114

FULL PRESCRIBING INFORMATION

WARNING: HEPATOTOXICITY  Severe and sometimes fatal hepatotoxicity has occurred in clinical trials [see Warnings and Precautions (5.1)].

 Monitor hepatic function prior to and during treatment [see Warnings and Precautions (5.1)].

 Interrupt and then reduce or discontinue STIVARGA for hepatotoxicity as manifested by elevated liver function tests or hepatocellular necrosis, depending upon severity and persistence [see Dosage and Administration (2.2)].

1 INDICATIONS AND USAGE 1.1 Colorectal Cancer STIVARGA is indicated for the treatment of patients with metastatic colorectal cancer (CRC) who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy, and, if RAS wild- type, an anti-EGFR therapy. 1.2 Gastrointestinal Stromal Tumors STIVARGA is indicated for the treatment of patients with locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate.

1.3 Hepatocellular Carcinoma STIVARGA is indicated for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with sorafenib. 2 DOSAGE AND ADMINISTRATION 2.1 Recommended Dose The recommended dose is 160 mg STIVARGA (four 40 mg tablets) taken orally once daily for the first 21 days of each 28-day cycle. Continue treatment until disease progression or unacceptable toxicity. Take STIVARGA at the same time each day. Swallow tablet whole with water after a low-fat meal that contains less than 600 calories and less than 30% fat [see Clinical Pharmacology (12.3)]. Do not take two doses of STIVARGA on the same day to make up for a missed dose from the previous day.

2.2 Dose Modifications If dose modifications are required, reduce the dose in 40 mg (one tablet) increments; the lowest recommended daily dose of STIVARGA is 80 mg daily. Interrupt STIVARGA for the following:  Grade 2 hand-foot skin reaction (HFSR) [palmar-plantar erythrodysesthesia syndrome (PPES)] that is recurrent or does not improve within 7 days despite dose reduction; interrupt therapy for a minimum of 7 days for Grade 3 HFSR  Symptomatic Grade 2 hypertension  Any Grade 3 or 4 adverse reaction  Worsening infection of any grade Reduce the dose of STIVARGA to 120 mg:  For the first occurrence of Grade 2 HFSR of any duration  After recovery of any Grade 3 or 4 adverse reaction except infection  For Grade 3 aspartate aminotransferase (AST)/alanine aminotransferase (ALT) elevation, only resume if the potential benefit outweighs the risk of hepatotoxicity

Reference ID: 4090114

Reduce the dose of STIVARGA to 80 mg:  For re-occurrence of Grade 2 HFSR at the 120 mg dose  After recovery of any Grade 3 or 4 adverse reaction at the 120 mg dose (except hepatotoxicity or infection) Discontinue STIVARGA permanently for the following:  Failure to tolerate 80 mg dose  Any occurrence of AST or ALT more than 20 times the upper limit of normal (ULN)  Any occurrence of AST or ALT more than 3 times ULN with concurrent bilirubin more than 2 times ULN  Re-occurrence of AST or ALT more than 5 times ULN despite dose reduction to 120 mg  For any Grade 4 adverse reaction; only resume if the potential benefit outweighs the risks 3 DOSAGE FORMS AND STRENGTHS STIVARGA is a 40 mg, light pink, oval-shaped, film-coated tablet, debossed with ‘BAYER’ on one side and ‘40’ on the other side. 4 CONTRAINDICATIONS None. 5 WARNINGS AND PRECAUTIONS 5.1 Hepatotoxicity Severe drug-induced liver injury with fatal outcome occurred in STIVARGA-treated patients in clinical trials. In most cases, liver dysfunction occurred within the first 2 months of therapy and was characterized by a hepatocellular pattern of injury. In the CORRECT study, fatal hepatic failure occurred in 1.6% of patients in the regorafenib arm and in 0.4% of patients in the placebo arm. In the GRID study, fatal hepatic failure occurred in 0.8% of patients in the regorafenib arm. In the RESORCE study, there was no increase in the incidence of fatal hepatic failure as compared to placebo [see Adverse Reactions (6.1)]. Obtain liver function tests (ALT, AST, and bilirubin) before initiation of STIVARGA and monitor at least every two weeks during the first 2 months of treatment. Thereafter, monitor monthly or more frequently as clinically indicated. Monitor liver function tests weekly in patients experiencing elevated liver function tests until improvement to less than 3 times the ULN or baseline. Temporarily hold and then reduce or permanently discontinue STIVARGA depending on the severity and persistence of hepatotoxicity as manifested by elevated liver function tests or hepatocellular necrosis [see Dosage and Administration (2.2) and Use in Specific Populations (8.6)]. 5.2 Infections STIVARGA caused an increased risk of infections. The overall incidence of infection (Grades 1-5) was higher (32% vs. 17%) in 1142 STIVARGA-treated patients as compared to the control arm in randomized placebo-controlled trials.The incidence of grade 3 or greater infections in STIVARGA treated patients was 9%. The most common infections were urinary tract infections (5.7%), nasopharyngitis (4.0%), mucocutaneous and systemic fungal infections (3.3%) and pneumonia (2.6%). Fatal outcomes caused by infection occurred more often in patients treated with STIVARGA (1.0%) as compared to patients receiving placebo (0.3%); the most common fatal infections were respiratory (0.6% in STIVARGA-treated patients vs 0.2% in patients receiving placebo). Withhold STIVARGA for Grade 3 or 4 infections, or worsening infection of any grade. Resume STIVARGA at the same dose following resolution of infection [see Dosage and Administration (2.2)]. 5.3 Hemorrhage STIVARGA caused an increased incidence of hemorrhage. The overall incidence (Grades 1-5) was 18.2% in 1142 patients treated with STIVARGA and 9.5% in patients receiving placebo in randomized, placebo-controlled trials. The incidence of grade 3 or greater hemorrhage in patients treated with STIVARGA was 3.0%. The incidence of fatal

Reference ID: 4090114

hemorrhagic events was 0.7%, involving the central nervous system or the respiratory, gastrointestinal, or genitourinary tracts. Permanently discontinue STIVARGA in patients with severe or life-threatening hemorrhage. Monitor INR levels more frequently in patients receiving warfarin [see Clinical Pharmacology (12.3)]. 5.4 Gastrointestinal Perforation or Fistula Gastrointestinal perforation occurred in 0.6% of 4518 patients treated with STIVARGA across all clinical trials of STIVARGA administered as as single agent; this included eight fatal events. Gastrointestinal fistula occurred in 0.8% of patients treated with STIVARGA and 0.2% of patients in placebo arm across randomized, placebo-controlled trials. Permanently discontinue STIVARGA in patients who develop gastrointestinal perforation or fistula. 5.5 Dermatologic Toxicity In randomized, placebo-controlled trials, adverse skin reactions occurred in 71.9% of patients in the regorafenib arm and in 25.5% of patients in the placebo arm, including hand-foot skin reaction (HFSR) also known as palmar-plantar erythrodysesthesia syndrome (PPES), and severe rash requiring dose modification. In the randomized, placebo-controlled trials, the overall incidence of HFSR was higher in 1142 STIVARGA-treated patients (53%) than in the placebo-treated patients (8%). Most cases of HFSR in STIVARGA-treated patients appeared during the first cycle of treatment. The incidences of Grade 3 HFSR (16% versus <1%), Grade 3 rash (3% versus <1%), serious adverse reactions of erythema multiforme (<0.1% vs. 0%) and Stevens-Johnson Syndrome (<0.1% vs. 0%) were also higher in STIVARGA-treated patients [see Adverse Reactions (6.1)]. Across all trials, a higher incidence of HFSR was observed in Asian patients treated with STIVARGA (all grades: 72%; Grade 3: 18%) [see Use in Specific Populations (8.8 )]. Toxic epidermal necrolysis occurred in 0.02% of 4518 STIVARGA-treated patients across all clinical trials of STIVARGA administered as a single agent. Withhold STIVARGA, reduce the dose, or permanently discontinue STIVARGA depending on the severity and persistence of dermatologic toxicity [see Dosage and Administration (2.2)]. Institute supportive measures for symptomatic relief. 5.6 Hypertension In randomized, placebo-controlled trials, hypertensive crisis occurred in 0.2% of patients in the regorafenib arms and in none of the patients in the placebo arms. STIVARGA caused an increased incidence of hypertension (30% versus 8% in CORRECT, 59% versus 27% in GRID, and 31% versus 6% in RESORCE) [see Adverse Reactions (6.1)]. The onset of hypertension occurred during the first cycle of treatment in most patients who developed hypertension (67% in randomized, placebo-controlled trials). Do not initiate STIVARGA unless blood pressure is adequately controlled. Monitor blood pressure weekly for the first 6 weeks of treatment and then every cycle, or more frequently, as clinically indicated. Temporarily or permanently withhold STIVARGA for severe or uncontrolled hypertension [see Dosage and Administration (2.2)]. 5.7 Cardiac Ischemia and Infarction STIVARGA increased the incidence of myocardial ischemia and infarction (0.9% vs 0.2%) in randomized placebo- controlled trials [see Adverse Reactions (6.1)]. Withhold STIVARGA in patients who develop new or acute onset cardiac ischemia or infarction. Resume STIVARGA only after resolution of acute cardiac ischemic events, if the potential benefits outweigh the risks of further cardiac ischemia. 5.8 Reversible Posterior Leukoencephalopathy Syndrome Reversible posterior leukoencephalopathy syndrome (RPLS), a syndrome of subcortical vasogenic edema diagnosed by characteristic finding on MRI, occurred in one of 4800 STIVARGA-treated patients across all clinical trials. Perform an evaluation for RPLS in any patient presenting with seizures, severe headache, visual disturbances, confusion or altered mental function. Discontinue STIVARGA in patients who develop RPLS.

Reference ID: 4090114

5.9 Wound Healing Complications No formal studies of the effect of regorafenib on wound healing have been conducted. Since vascular endothelial growth factor receptor (VEGFR) inhibitors such as STIVARGA can impair wound healing, discontinue treatment with STIVARGA at least 2 weeks prior to scheduled surgery. The decision to resume STIVARGA after surgery should be based on clinical judgment of adequate wound healing. Discontinue STIVARGA in patients with wound dehiscence. 5.10 Embryo-Fetal Toxicity Based on animal studies and its mechanism of action, STIVARGA can cause fetal harm when administered to a pregnant woman. There are no available data on STIVARGA use in pregnant women. Regorafenib was embryolethal and teratogenic in rats and rabbits at exposures lower than human exposures at the recommended dose, with increased incidences of cardiovascular, genitourinary, and skeletal malformations. Advise pregnant women of the potential risk to a fetus. Advise females of reproductive potential to use effective contraception during treatment with STIVARGA and for 2 months after the final dose. Advise males with female partners of reproductive potential to use effective contraception during treatment with STIVARGA and for 2 months after the final dose [see Use in Specific Populations (8.1), (8.3)]. 6 ADVERSE REACTIONS The following serious adverse reactions are discussed elsewhere in the labeling:  Hepatotoxicity [see Warnings and Precautions (5.1)]  Infections [(see Warnings and Precautions (5.2)]  Hemorrhage [see Warnings and Precautions (5.3)]  Gastrointestinal Perforation or Fistula [see Warnings and Precautions (5.4)]  Dermatological Toxicity [see Warnings and Precautions (5.5)]  Hypertension [see Warnings and Precautions (5.6)]  Cardiac Ischemia and Infarction [see Warnings and Precautions (5.7)]  Reversible Posterior Leukoencephalopathy Syndrome (RPLS) [see Warnings and Precautions (5.8)] 6.1 Clinical Trials Experience Because clinical trials are conducted under widely varying conditions, adverse reaction rates observed in the clinical trials of a drug cannot be directly compared to rates in the clinical trials of another drug and may not reflect the rate observed in practice. The data described in the WARNINGS AND PRECAUTIONS section reflect exposure to STIVARGA in more than 4800 patients who were enrolled in four randomized, placebo-controlled trials (n=1142), an expanded access program (CONSIGN, n=2864), or single arm clinical trials (single agent or in combination with other agents). There were 4518 patients who received STIVARGA as a single agent; the distribution of underlying malignancies was 80% CRC, 4% GIST, 10% HCC, 6% other solid tumors; and 74% were White, 11% Asian, and 15% race not known. Among these 4518 patients, 83% received STIVARGA for at least 21 days and 20% received STIVARGA for 6 months or longer. In randomized placebo-controlled trials (CORRECT, GRID, RESORCE and CONCUR), the most frequently observed adverse drug reactions (≥20%) in patients receiving STIVARGA are pain (including gastrointestinal and abdominal pain), HFSR, asthenia/fatigue, diarrhea, decreased appetite/food intake, hypertension, infection, dysphonia, hyperbilirubinemia, fever, mucositis, weight loss, rash, and nausea.

Reference ID: 4090114

Colorectal Cancer The safety data described below, except where noted, are derived from a randomized (2:1), double-blind, placebo- controlled trial (CORRECT) in which 500 patients (median age 61 years; 61% men) with previously-treated metastatic colorectal cancer (CRC) received STIVARGA as a single agent at the dose of 160 mg daily for the first 3 weeks of each 4 week treatment cycle and 253 patients (median age 61 years; 60% men) received placebo. The median duration of therapy was 1.7 months (range 2 days, 10.8 months) for patients receiving STIVARGA. Due to adverse reactions, 61% of the patients receiving STIVARGA required a dose interruption and 38% of the patients had their dose reduced. Adverse reactions that resulted in treatment discontinuation occurred in 8.2% of STIVARGA-treated patients compared to 1.2% of patients who received placebo. Hand-foot skin reaction (HFSR) and rash were the most common reasons for permanent discontinuation of STIVARGA. Table 1 provides the incidence of adverse reactions (≥10%) in patients in CORRECT. Table 1: Adverse drug reactions reported in ≥10% of patients treated with STIVARGA in CORRECT and reported more commonly than in patients receiving placeboa

STIVARGA Placebo (N=500) (N=253) Adverse Reactions Grade Grade All ≥ 3 All ≥ 3 % % % % General disorders and administration site conditions Asthenia/fatigue 64 15 46 9 Pain 59 9 48 7 Fever 28 2 15 0 Metabolism and nutrition disorders Decreased appetite and food intake 47 5 28 4 Skin and subcutaneous tissue disorders HFSR/PPES 45 17 7 0 Rash b 26 6 4 <1 Gastrointestinal disorders Diarrhea 43 8 17 2 Mucositis 33 4 5 0 Investigations Weight loss 32 <1 10 0 Infections and infestations Infection c 31 9 17 6 Vascular disorders Hypertension 30 8 8 <1 Hemorrhage c 21 2 8 <1 Respiratory, thoracic and mediastinal disorders Dysphonia 30 0 6 0 Nervous system disorders Headache 10 <1 7 0 a Adverse reactions graded according to National Cancer Institute Common Toxicity for Adverse Events version 3.0 (NCI CTCAE v3.0). b The term rash represents reports of events of drug eruption, rash, erythematous rash, generalized rash, macular rash, maculo-papular rash, papular rash, and pruritic rash. c Fatal outcomes observed.

Reference ID: 4090114

Table 2 provides laboratory abnormalities observed in CORRECT. Table 2: Laboratory test abnormalities reported in CORRECT

STIVARGA Placebo (N=500 a) (N=253 a) b b Laboratory Parameter Grade Grade All 3 4 All 3 4 % % % % % % Blood and lymphatic system disorders Anemia 79 5 1 66 3 0 Thrombocytopenia 41 2 <1 17 <1 0 Neutropenia 3 1 0 0 0 0 Lymphopenia 54 9 0 35 4 <1 Metabolism and nutrition disorders Hypocalcemia 59 1 <1 18 1 0 Hypokalemia 26 4 0 8 <1 0 Hyponatremia 30 7 1 22 4 0 Hypophosphatemia 57 31 1 11 4 0 Hepatobiliary disorders Hyperbilirubinemia 45 10 3 17 5 3 Increased AST 65 5 1 46 4 1 Increased ALT 45 5 1 30 3 <1 Renal and urinary disorders Proteinuriac 84 2 0 61 1 0 Investigations Increased INRd 24 4 N/A 17 2 N/A Increased Lipase 46 9 2 19 3 2 Increased Amylase 26 2 <1 17 2 <1 a % based on number of patients with post-baseline samples which may be less than 500 (regorafenib) or 253 (placebo). b NCI CTCAE v3.0. c Based on urine protein-creatinine ratio data. d International normalized ratio: No Grade 4 denoted in NCI CTCAE, v3.0. Gastrointestinal Stromal Tumors The safety data described below are derived from a randomized (2:1), double-blind, placebo-controlled trial (GRID) in which 132 patients (median age 60 years; 64% men) with previously-treated GIST received STIVARGA as a single agent at a dose of 160 mg daily for the first 3 weeks of each 4 week treatment cycle and 66 patients (median age 61 years; 64% men) received placebo. The median duration of therapy was 5.7 months (range 1 day, 11.7 months) for patients receiving STIVARGA. Dose interruptions for adverse events were required in 58% of patients receiving STIVARGA and 50% of patients had their dose reduced. Adverse reactions that resulted in treatment discontinuation were reported in 2.3% of STIVARGA-treated patients compared to 1.5% of patients who received placebo. Table 3 provides the incidence of adverse reactions (≥10%) in patients in GRID.

Reference ID: 4090114

Table 3: Adverse reactions reported in ≥10% patients treated with STIVARGA in GRID and reported more commonly than in patients receiving placeboa

STIVARGA Placebo (N=132) (N=66) Adverse Reactions Grade Grade All ≥ 3 All ≥ 3 % % % % Skin and subcutaneous tissue disorders HFSR/PPE 67 22 12 2 Rash b 30 7 3 0 Alopecia 24 2 2 0 General disorders and administration site conditions Asthenia/Fatigue 52 4 39 2 Fever 21 0 11 2 Vascular disorders Hypertension 59 28 27 5 Hemorrhage 11 4 3 0 Gastrointestinal disorders Pain 60 8 55 14 Diarrhea 47 8 9 0 Mucositis 40 2 8 2 Nausea 20 2 12 2 Vomiting 17 <1 8 0 Respiratory, thoracic and mediastinal disorders Dysphonia 39 0 9 0 Infections and infestations Infection c 32 5 5 0 Metabolism and nutrition disorders Decreased appetite and food intake 31 <1 21 3 Hypothyroidism d 18 0 6 0 Nervous system disorders Headache 16 0 9 0 Investigations Weight loss 14 0 8 0 Musculoskeletal and connective tissue disorders Muscle spasms 14 0 3 0 a Adverse reactions graded according to NCI CTCAE v4.0. b The term rash represents reports of events of rash, erythematous rash, macular rash, maculo-papular rash, papular rash and pruritic rash. c Fatal outcomes observed. d Hypothyroidism incidence based on subset of patients with normal TSH and no thyroid supplementation at baseline.

Table 4 provides laboratory abnormalities observed in GRID.

Reference ID: 4090114

Table 4: Laboratory test abnormalities reported in GRID

STIVARGA Placebo (N=132 a) (N=66 a) Laboratory Parameter Grade b Grade b All 3 4 All 3 4 % % % % % % Blood and lymphatic system disorders Thrombocytopenia 13 1 0 2 0 2 Neutropenia 16 2 1 12 3 0 Lymphopenia 30 8 0 24 3 0 Metabolism and nutrition disorders Hypocalcemia 17 2 0 5 0 0 Hypokalemia 21 3 0 3 0 0 Hypophosphatemia 55 20 2 3 2 0 Hepatobiliary disorders Hyperbilirubinemia 33 3 1 12 2 0 Increased AST 58 3 1 47 3 0 Increased ALT 39 4 1 39 2 0 Renal and urinary disorders Proteinuria c 59 3 - d 53 3 - d Investigations Increased Lipase 14 0 1 5 0 0 a Percent based on number of patients with post-baseline samples which may be less than 132 (regorafenib) or 66 (placebo). b NCI CTCAE v4.0. c Based on urine protein-creatinine ratio data. d No Grade 4 denoted in NCI CTCAE v4.0.

Hepatocellular Carcinoma

The safety data described below are derived from a randomized (2:1), double-blind, placebo-controlled trial (RESORCE) in which patients with previously-treated HCC received either STIVARGA (n=374) 160 mg orally on days 1-21 of each 4 week treatment cycle or placebo (n=193). The median age was 63 years, 88% were men, 98% had Child-Pugh A cirrhosis, 66% had an ECOG performance status (PS) of 0 and 34% had PS of 1. The median duration of therapy was 3.5 months (range 1 day to 29.4 months) for patients receiving STIVARGA. Of the patients receiving STIVARGA, 33% were exposed to STIVARGA for greater than or equal to 6 months and 14% were exposed to STIVARGA for greater than or equal to 12 months. Dose interruptions for adverse events were required in 58.3% of patients receiving STIVARGA and 48% of patients had their dose reduced. The most common adverse reactions requiring dose modification (interruption or dose reduction) were HFSR/PPES (20.6%), blood bilirubin increase (5.9%), fatigue (5.1%) and diarrhea (5.3%). Adverse reactions that resulted in treatment discontinuation were reported in 10.4% of STIVARGA-treated patients compared to 3.6% of patients who received placebo; the most common adverse reactions requiring discontinuation of STIVARGA were HFSR/PPES (1.9%) and AST increased (1.6%). Table 5 provides the incidence of adverse reactions (≥10%) in patients in RESORCE.

Reference ID: 4090114

Table 5: Adverse reactions reported in ≥10% of patients treated with STIVARGA in RESORCE and reported more commonly than in patients receiving placeboa

STIVARGA Placebo (N=374) (N=193) Adverse Reactions Grade Grade All ≥ 3 All ≥ 3 % % % % Skin and subcutaneous tissue disorders HFSR/PPE 51 12 7 <1 General disorders and administration site conditions Pain 55 9 44 8 Asthenia/Fatigue 42 10 33 5 Fever 20 0 7 0 Vascular disorders Hypertension 31 15 6 5 Hemorrhage b 18 5 16 8 Gastrointestinal disorders Diarrhea 41 3 15 0 Nausea 17 <1 13 0 Vomiting 13 <1 7 <1 Mucositis 13 1 2 ≤1 Respiratory, thoracic and mediastinal disorders Dysphonia 18 0 2 0 Infections and infestations Infection b 31 8 18 6 Metabolism and nutrition disorders Decreased appetite and food intake 31 3 15 2 Investigations Weight loss 13 2 4 0

Musculoskeletal and connective tissue disorders Muscle spasms 10 0 2 0 a Adverse reactions graded according to NCI CTCAE v4.0. b Fatal outcomes observed.

Other clinically significant adverse reactions observed in less than 10% of STIVARGA-treated patients were: alopecia (7%), hypothyroidism (6.4%), pancreatitis (1.6%), exfoliative rash (1.3%), tremor (1.3%), erythema multiforme (0.8%), myocardial ischemia (0.8%), gastrointestinal fistula (0.3%), and myocardial infarction (0.3%).

Table 6 provides laboratory abnormalities observed in RESORCE.

Reference ID: 4090114

Table 6: Laboratory test abnormalities reported in RESORCE

STIVARGA Placebo (N=374 a) (N=193 a) Laboratory Parameter Grade b Grade b All 3 4 All 3 4 % % % % % % Blood and lymphatic system disorders Thrombocytopenia 63 5 <1 50 0 0 Neutropenia 14 3 0 15 <1 <1 Lymphopenia 68 16 2 59 11 <1 Metabolism and nutrition disorders Hypocalcemia 23 <1 0 10 0 0 Hypokalemia 31 4 <1 9 2 0 Hypophosphatemia 70 32 2 31 7 0 Hepatobiliary disorders Hyperbilirubinemia 78 13 3 55 11 5 Increased AST 93 16 2 84 17 3 Increased ALT 70 6 <1 59 5 0 Renal and urinary disorders Proteinuria c 51 17 - d 37 3 - d Investigations Increased INR 44 <1 - d 35 2 - d Increased Lipase 41 11 3 27 8 1 Increased Amylase 23 3 <1 19 2 <1 a Percent based on number of patients with post-baseline samples which may be less than 374 (regorafenib) or 193 (placebo). b NCI CTCAE v4.0. c Based on dipstick data. d No Grade 4 denoted in NCI CTCAE v4.0.

6.2 Postmarketing Experience The following adverse reaction has been identified during postapproval use of STIVARGA. Because these reactions are reported voluntarily from a population of uncertain size, it is not always possible to reliably estimate their frequency or establish a causal relationship to drug exposure:

 hypersensitivity reaction 7 DRUG INTERACTIONS 7.1 Effect of Strong CYP3A4 Inducers on Regorafenib Co-administration of a strong CYP3A4 inducer with STIVARGA decreased the plasma concentrations of regorafenib, increased the plasma concentrations of the active metabolite M-5, and resulted in no change in the plasma concentrations of the active metabolite M-2 [see Clinical Pharmacology (12.3)], and may lead to decreased efficacy. Avoid concomitant use of STIVARGA with strong CYP3A4 inducers (e.g. rifampin, phenytoin, carbamazepine, phenobarbital, and St. John’s Wort).

Reference ID: 4090114

7.2 Effect of Strong CYP3A4 Inhibitors on Regorafenib Co-administration of a strong CYP3A4 inhibitor with STIVARGA increased the plasma concentrations of regorafenib and decreased the plasma concentrations of the active metabolites M-2 and M-5 [see Clinical Pharmacology (12.3)], and may lead to increased toxicity. Avoid concomitant use of STIVARGA with strong CYP3A4 inhibitors (e.g. clarithromycin, grapefruit juice, itraconazole, ketoconazole, nefazodone, posaconazole, telithromycin, and voriconazole). 7.3 Effect of Regorafenib on Breast Cancer Resistance Protein (BCRP) Substrates Co-administration of STIVARGA with a BCRP substrate increased the plasma concentrations of the BCRP substrate [see Clinical Pharmacology (12.3)]. Monitor patients closely for signs and symptoms of exposure related toxicity to the BCRP substrate (e.g. methotrexate, fluvastatin, atorvastatin). Consult the concomitant BCRP substrate product information when considering administration of such products together with STIVARGA. 8 USE IN SPECIFIC POPULATIONS 8.1 Pregnancy Risk Summary Based on animal studies and its mechanism of action, STIVARGA can cause fetal harm when administered to a pregnant woman. There are no available data on STIVARGA use in pregnant women. Administration of regorafenib was embryolethal and teratogenic in rats and rabbits at exposures lower than human exposures at the recommended dose, with increased incidences of cardiovascular, genitourinary, and skeletal malformations [see Data]. Advise pregnant women of the potential hazard to a fetus. The estimated background risk of major birth defects and miscarriage for the indicated population is unknown. In the U.S. general population, the estimated background risk of major birth defects and miscarriage in clinically recognized pregnancies is 2 to 4 % and 15 to 20%, respectively. Data Animal Data In embryo-fetal development studies, a total loss of pregnancy (100% resorption of litter) was observed in rats at doses as low as 1 mg/kg (approximately 6% of the recommended human dose, based on body surface area) and in rabbits at doses as low as 1.6 mg/kg (approximately 25% of the human exposure at the clinically recommended dose measured by AUC). In a single dose distribution study in pregnant rats, there was increased penetration of regorafenib across the blood-brain barrier in fetuses compared to dams. Daily administration of regorafenib to pregnant rats during organogenesis resulted in fetal findings of delayed ossification at doses > 0.8 mg/kg (approximately 5% of the recommended human dose based on body surface area) and dose-dependent increases in skeletal malformations including cleft palate and enlarged fontanelle at doses ≥ 1 mg/kg (approximately 10% of the clinical exposure based on AUC). At doses ≥ 1.6 mg/kg (approximately 11% of the recommended human dose based on body surface area), there were dose-dependent increases in the incidence of cardiovascular malformations, external abnormalities, diaphragmatic hernia, and dilation of the renal pelvis. In pregnant rabbits administered regorafenib daily during organogenesis, there were findings of ventricular septal defects evident at the lowest tested dose of 0.4 mg/kg (approximately 7% of the AUC in patients at the recommended dose). At doses of ≥ 0.8 mg/kg (approximately 15% of the human exposure at the recommended human dose based on AUC), administration of regorafenib resulted in dose-dependent increases in the incidence of additional cardiovascular malformations and skeletal anomalies, as well as significant adverse effects on the urinary system including missing kidney/ureter; small, deformed and malpositioned kidney; and hydronephrosis. The proportion of viable fetuses that were male decreased with increasing dose in two rabbit embryo-fetal toxicity studies.

Reference ID: 4090114

8.2 Lactation Risk Summary There are no data on the presence of regorafenib or its metabolites in human milk, the effects of regorafenib on the breastfed infant, or on milk production. In rats, regorafenib and its metabolites are excreted in milk. Because of the potential for serious adverse reactions in breastfed infants from STIVARGA, do not breastfeed during treatment with STIVARGA and for 2 weeks after the final dose. 8.3 Females and Males of Reproductive Potential Contraception Females Use effective contraception during treatment and for 2 months after completion of therapy. Males Advise male patients with female partners of reproductive potential to use effective contraception during treatment and for 2 months following the final dose of STIVARGA [see Nonclinical Toxicology (13.1)]. Infertility There are no data on the effect of STIVARGA on human fertility. Results from animal studies indicate that regorafenib can impair male and female fertility [see Nonclinical Toxicology (13.1)]. 8.4 Pediatric Use The safety and efficacy of STIVARGA in pediatric patients less than 18 years of age have not been established. Animal Data In 28-day repeat-dose studies in rats there were dose-dependent findings of dentin alteration and angiectasis. These findings occurred at regorafenib doses as low as 4 mg/kg (approximately 25% of the AUC in humans at the recommended dose). In 13-week repeat-dose studies in dogs there were similar findings of dentin alteration at doses as low as 20 mg/kg (approximately 43% of the AUC in humans at the recommended dose). Administration of regorafenib in these animals also led to persistent growth and thickening of the femoral epiphyseal growth plate. 8.5 Geriatric Use Of the 1142 STIVARGA-treated patients enrolled in randomized, placebo-controlled trials, 40% were 65 years of age and over, while 10% were 75 and over. No overall differences in efficacy were observed between these patients and younger patients. There was an increased incidence of Grade 3 hypertension (18% versus 9%) in the placebo-controlled trials among STIVARGA-treated patients 65 years of age and older as compared to younger patients. In addition, one Grade 4 hypertension event has been reported in the 65 years and older age group and none in the younger age group. 8.6 Hepatic Impairment No dose adjustment is recommended in patients with mild (total bilirubin ULN and AST >ULN, or total bilirubin >ULN to ≤1.5 times ULN) or moderate (total bilirubin >1.5 to ≤3 times ULN and any AST) hepatic impairment, [see Clinical Pharmacology (12.3)]. Closely monitor patients with hepatic impairment for adverse reactions [see Warnings and Precautions (5.1)]. STIVARGA is not recommended for use in patients with severe hepatic impairment (total bilirubin >3x ULN) as STIVARGA has not been studied in this population. 8.7 Renal Impairment No dose adjustment is recommended for patients with renal impairment. The pharmacokinetics of regorafenib have not been studied in patients who are on dialysis and there is no recommended dose for this patient population [see Clinical Pharmacology (12.3)].

Reference ID: 4090114

8.8 Race Based on pooled data from three placebo-controlled trials (CORRECT, GRID and CONCUR), a higher incidence of HFSR and liver function test abnormalities occurred in Asian patients treated with STIVARGA as compared with Whites [see Warnings and Precautions (5.1, 5.5)]. No starting dose adjustment is necessary based on race. 10 OVERDOSAGE The highest dose of STIVARGA studied clinically is 220 mg per day. The most frequently observed adverse drug reactions at this dose were dermatological events, dysphonia, diarrhea, mucosal inflammation, dry mouth, decreased appetite, hypertension, and fatigue. There is no known antidote for STIVARGA overdose. In the event of suspected overdose, interrupt STIVARGA, institute supportive care, and observe until clinical stabilization. 11 DESCRIPTION STIVARGA (regorafenib) is a multikinase inhibitor with the chemical name 4-[4-({[4-chloro-3-(trifluoromethyl) phenyl] carbamoyl} amino)-3-fluorophenoxy]-N-methylpyridine-2-carboxamide monohydrate. Regorafenib has the following structural formula:

Regorafenib is a monohydrate and it has a molecular formula C21H15ClF4N4O3 • H2O and a molecular weight of 500.83. Regorafenib is practically insoluble in water, slightly soluble in acetonitrile, methanol, ethanol, and ethyl acetate and sparingly soluble in acetone. STIVARGA tablets for oral administration are formulated as light pink, oval-shaped tablets debossed with "BAYER" on one side and "40" on the other. Each tablet contains 40 mg of regorafenib in the anhydrous state, which corresponds to 41.49 mg of regorafenib monohydrate, and the following inactive ingredients: cellulose microcrystalline, croscarmellose sodium, magnesium stearate, povidone, and colloidal silicon dioxide. The film-coating contains the following inactive ingredients: ferric oxide red, ferric oxide yellow, lecithin (soy), polyethylene glycol 3350, polyvinyl alcohol, talc, and titanium dioxide. 12 CLINICAL PHARMACOLOGY 12.1 Mechanism of Action Regorafenib is a small molecule inhibitor of multiple membrane-bound and intracellular kinases involved in normal cellular functions and in pathologic processes such as oncogenesis, tumor angiogenesis, metastasis and tumor immunity. In in vitro biochemical or cellular assays, regorafenib or its major human active metabolites M-2 and M-5 inhibited the activity of RET, VEGFR1, VEGFR2, VEGFR3, KIT, PDGFR-alpha, PDGFR-beta, FGFR1, FGFR2, TIE2, DDR2, TrkA, Eph2A, RAF-1, BRAF, BRAF V600E, SAPK2, PTK5, Abl and CSF1R at concentrations of regorafenib that have been achieved clinically. In in vivo models, regorafenib demonstrated anti-angiogenic activity in a rat tumor model and inhibition of tumor growth in several mouse xenograft models including some for human colorectal carcinoma, gastrointestinal stromal and hepatocellular carcinoma. Regorafenib also demonstrated anti-metastatic activity in a mouse xenograft model and two mouse orthotopic models of human colorectal carcinoma.

Reference ID: 4090114

12.2 Pharmacodynamics Cardiac Electrophysiology The effect of multiple doses of STIVARGA (160 mg once daily for 21 days) on the QTc interval was evaluated in an open-label, single-arm study in 25 patients with advanced solid tumors. No large changes in the mean QTc interval (i.e., > 20 msec) were detected in the study. 12.3 Pharmacokinetics Absorption Following a single 160 mg dose of STIVARGA in patients with advanced solid tumors, regorafenib reaches a geometric mean peak plasma level (Cmax) of 2.5 µg/mL at a median time of 4 hours and a geometric mean area under the plasma concentration vs. time curve (AUC) of 70.4 µg*h/mL. The AUC of regorafenib at steady-state increases less than dose proportionally at doses greater than 60 mg. At steady-state, regorafenib reaches a geometric mean Cmax of 3.9 µg/mL and a geometric mean AUC of 58.3 µg*h/mL. The coefficient of variation of AUC and Cmax is between 35% and 44%. The mean relative bioavailability of tablets compared to an oral solution is 69% to 83%. In a food-effect study, 24 healthy men received a single 160 mg dose of STIVARGA on three separate occasions: under a fasted state, with a high-fat meal and with a low-fat meal. A high-fat meal (945 calories and 54.6 g fat) increased the mean AUC of regorafenib by 48% and decreased the mean AUC of the M-2 and M-5 metabolites by 20% and 51%, respectively, as compared to the fasted state. A low-fat meal (319 calories and 8.2 g fat) increased the mean AUC of regorafenib, M-2 and M-5 by 36%, 40% and 23%, respectively as compared to fasted conditions. STIVARGA was administered with a low-fat meal in the CORRECT and GRID studies [see Dosage and Administration (2.1), Clinical Studies (14)]. Distribution Regorafenib undergoes enterohepatic circulation with multiple plasma concentration peaks observed across the 24-hour dosing interval. Regorafenib is highly bound (99.5%) to human plasma proteins. Elimination Following a single 160 mg oral dose of STIVARGA, the geometric mean (minimum to maximum) elimination half-lives for regorafenib and the M-2 metabolite in plasma are 28 hours (14 to 58 hours) and 25 hours (14 to 32 hours), respectively. M-5 has a longer mean (minimum to maximum) elimination half-life of 51 hours (32 to 70 hours). Metabolism Regorafenib is metabolized by CYP3A4 and UGT1A9. The main circulating metabolites of regorafenib measured at steady-state in human plasma are M-2 (N-oxide) and M-5 (N-oxide and N-desmethyl). Both metabolites have similar in vitro pharmacological activity and steady-state concentrations as regorafenib. M-2 and M-5 are highly protein bound (99.8% and 99.95%, respectively). Excretion Approximately 71% of a radiolabeled dose was excreted in feces (47% as parent compound, 24% as metabolites) and 19% of the dose was excreted in urine (17% as glucuronides) within 12 days after administration of a radiolabeled oral solution at a dose of 120 mg. Specific Populations Age, sex, race and weight had no clinically meaningful effect on the pharmacokinetics of regorafenib. Hepatic Impairment Based on a population pharmacokinetic analysis, no clinically important differences in the mean total exposure of regorafenib, including M-2 and M-5, were noted amongst patients with normal liver function (total bilirubin and AST  ULN, n=744), mild hepatic impairment (total bilirubin  ULN and AST >ULN or total bilirubin >ULN to ≤1.5x ULN, n=437), and moderate hepatic impairment (total bilirubin >1.5x to ≤3x ULN and any AST, n=36). The pooled analysis included 391 patients with HCC of whom 116, 249, and 26 were categorized as having normal liver function, mild, and 16

Reference ID: 4090114

moderate hepatic impairment, respectively. The pharmacokinetics of regorafenib were not evaluated in patients with severe hepatic impairment (total bilirubin >3x ULN). Renal Impairment The pharmacokinetics of regorafenib, M-2, and M-5 was evaluated in 6 patients with severe renal impairment (CLcr 15 29 mL/min) and 18 patients with normal/mild renal function (CLcr ≥60 mL/min) following the administration of STIVARGA at a dose of 160 mg daily for 21 days. No differences in the mean steady-state exposure of regorafenib, M-2, or M-5 were observed in patients with severe renal impairment compared to patients with normal renal function. The pharmacokinetics of regorafenib has not been studied in patients with end-stage renal disease on dialysis. Drug Interaction Studies Effect of Regorafenib on Cytochrome P450 Substrates: In vitro studies suggested that regorafenib is an inhibitor of CYP2C8, CYP2C9, CYP2B6, CYP3A4 and CYP2C19; M-2 is an inhibitor of CYP2C9, CYP2C8, CYP3A4 and CYP2D6, and M-5 is an inhibitor of CYP2C8. In vitro studies suggested that regorafenib is not an inducer of CYP1A2, CYP2B6, CYP2C19, and CYP3A4 enzyme activity. Patients with advanced solid tumors received single oral doses of CYP substrates, 2 mg of midazolam (CYP3A4), 40 mg of omeprazole (CYP2C19) and 10 mg of warfarin (CYP2C9) or 4 mg of rosiglitazone (CYP2C8) one week before and two weeks after STIVARGA at a dose of 160 mg once daily. No clinically meaningful effect was observed in the mean AUC of rosiglitazone (N=12) or the mean omeprazole (N=11) plasma concentrations measured 6 hours after dosing or the mean AUC of midazolam (N=15). The mean AUC of warfarin (N=8) increased by 25% [see Warnings and Precautions (5.2)]. Effect of CYP3A4 Strong Inducers on Regorafenib: Twenty-two healthy men received a single 160 mg dose of STIVARGA alone and then 7 days after starting rifampin. Rifampin, a strong CYP3A4 inducer, was administered at a dose of 600 mg daily for 9 days. The mean AUC of regorafenib decreased by 50% and mean AUC of M-5 increased by 264%. No change in the mean AUC of M-2 was observed [see Drug Interactions (7.1)]. Effect of CYP3A4 Strong Inhibitors on Regorafenib: Eighteen healthy men received a single 160 mg dose of STIVARGA alone and then 5 days after starting ketoconazole. Ketoconazole, a strong CYP3A4 inhibitor, was administered at a dose of 400 mg daily for 18 days. The mean AUC of regorafenib increased by 33% and the mean AUC of M-2 and M-5 both decreased by 93% [see Drug Interactions (7.2)]. Effect of Neomycin on Regorafenib: Twenty-seven healthy men received a single 160 mg dose of STIVARGA and then 5 days after starting neomycin. Neomycin, a non-absorbable antibiotic, was administered at a dose of 1 gram three times daily for 5 days. No clinically meaningful effect on the mean AUC of regorafenib was observed; however, the mean AUC of M-2 decreased by 76% and the mean AUC of M-5 decreased by 86%. The decreased exposure of M-2 and M-5 may result in a decreased efficacy of STIVARGA. The effects of other antibiotics on the exposure of regorafenib and its active metabolites have not been studied. Effect of Regorafenib on UGT1A1 Substrates: In vitro studies showed that regorafenib, M-2, and M-5 competitively inhibit UGT1A9 and UGT1A1 at therapeutically relevant concentrations. Eleven patients received irinotecan-containing combination chemotherapy with STIVARGA at a dose of 160 mg. The mean AUC of irinotecan increased by 28% and the mean AUC of SN-38 increased by 44% when irinotecan was administered 5 days after the last of 7 daily doses of STIVARGA. Effect of Regorafenib on BCRP Substrates: Administration of regorafenib (160 mg for 14 days) prior to administration of a single dose of rosuvastatin (5 mg), a BCRP substrate, resulted in a 3.8-fold increase in mean exposure (AUC) of rosuvastatin and a 4.6-fold increase in Cmax [see Drug Interactions (7.3)]. 13 NONCLINICAL TOXICOLOGY 13.1 Carcinogenesis, Mutagenesis, Impairment of Fertility Studies examining the carcinogenic potential of regorafenib have not been conducted. Regorafenib itself did not demonstrate genotoxicity in in vitro or in vivo assays; however, a major human active metabolite of regorafenib, (M-2), was positive for clastogenicity, causing chromosome aberration in Chinese hamster V79 cells.

Reference ID: 4090114

Dedicated studies to examine the effects of regorafenib on fertility have not been conducted; however, there were histological findings of tubular atrophy and degeneration in the testes, atrophy in the seminal vesicle, and cellular debris and oligospermia in the epididymides in male rats at doses similar to those in human at the clinical recommended dose based on AUC. In female rats, there were increased findings of necrotic corpora lutea in the ovaries at the same exposures. There were similar findings in dogs of both sexes in repeat dose studies at exposures approximately 83% of the human exposure at the recommended human dose based on AUC. These findings suggest that regorafenib may adversely affect fertility in humans. 13.2 Animal Toxicology and/or Pharmacology In a chronic 26-week repeat dose study in rats there was a dose-dependent increase in the finding of thickening of the atrioventricular valve. At a dose that resulted in an exposure of approximately 12% of the human exposure at the recommended dose, this finding was present in half of the examined animals. 14 CLINICAL STUDIES 14.1 Colorectal Cancer The clinical efficacy and safety of STIVARGA were evaluated in an international, multicenter, randomized (2:1), double- blind, placebo-controlled trial [Study “Patients with metastatic COloRectal cancer treated with REgorafenib or plaCebo after failure of standard Therapy” (CORRECT); NCT 01103323)] in 760 patients with previously-treated metastatic colorectal cancer. The major efficacy outcome measure was overall survival (OS); additional efficacy outcome measures included progression-free survival (PFS) and overall tumor response rate. Patients were randomized to receive 160 mg regorafenib orally once daily (N=505) plus best supportive care (BSC) or placebo (N=255) plus BSC for the first 21 days of each 28-day cycle. STIVARGA was administered with a low-fat breakfast that contains less than 30% fat [see Dosage and Administration (2.1), Clinical Pharmacology (12.3)]. Treatment continued until disease progression or unacceptable toxicity. Baseline demographics were: median age 61 years, 61% men, 78% White, and all patients had an ECOG performance status of 0 or 1. The primary sites of disease were colon (65%), rectum (29%), or both (6%). History of KRAS evaluation was reported for 729 (96%) patients; 430 (59%) of these patients were reported to have KRAS mutation. The median number of prior lines of therapy for metastatic disease was 3. All patients received prior treatment with fluoropyrimidine-, oxaliplatin-, and irinotecan-based chemotherapy, and with bevacizumab. All but one patient with KRAS mutation- negative tumors received panitumumab or cetuximab. The addition of STIVARGA to BSC resulted in a statistically significant improvement in survival compared to placebo plus BSC (see Table 7 and Figure 1).

Reference ID: 4090114

Table 7: Efficacy Results from CORRECT

STIVARGA Placebo (N=505) (N=255) Overall Survival Number of Deaths (%) 275 (55%) 157 (62%) Median Overall Survival (months) 6.4 5.0 95% CI a (5.8, 7.3) (4.4, 5.8) HR (95% CI) 0.77 (0.64, 0.94) Stratified log-rank test p-value b, c 0.0102 Progression-Free Survival Number of Deaths or Progressions (%) 417 (83%) 231 (91%) Median Progression-Free Survival (months) 2.0 1.7 95% CI (1.9, 2.3) (1.7, 1.8) HR (95% CI) 0.49 (0.42, 0.58) Stratified log-rank test p-value c <0.0001 Overall Response Rate Overall Response, N (%) 5 (1%) 1 (0.4%) 95% CI 0.3%, 2.3% 0%, 2.2% a CI=confidence interval. b Stratified by geographic region and time from diagnosis of metastatic disease. c Crossed the O’Brien-Fleming boundary (two-sided p-value < 0.018) at second interim analysis.

Figure 1: Kaplan-Meier Curves of Overall Survival

Reference ID: 4090114

14.2 Gastrointestinal Stromal Tumors The efficacy and safety of STIVARGA were evaluated in an international, multicenter, randomized (2:1), double-blind, placebo-controlled trial [Study “GIST Regorafenib In progressive Disease” (GRID); NCT 01271712] in 199 patients with unresectable, locally advanced or metastatic gastrointestinal stromal tumor (GIST), who had been previously treated with imatinib mesylate and sunitinib malate. Randomization was stratified by line of therapy (third vs. four or more) and geographic region (Asia vs. rest of the world). The major efficacy outcome measure of GRID was progression-free survival (PFS) based on disease assessment by independent radiological review using modified RECIST 1.1 criteria, in which lymph nodes and bone lesions were not target lesions and progressively growing new tumor nodule within a pre-existing tumor mass was progression. The key secondary outcome measure was overall survival. Patients were randomized to receive 160 mg regorafenib orally once daily (N=133) plus best supportive care (BSC) or placebo (N=66) plus BSC for the first 21 days of each 28-day cycle. Treatment continued until disease progression or unacceptable toxicity. In GRID, the median age of patients was 60 years, 64% were men, 68% were White, and all patients had baseline ECOG performance status of 0 (55%) or 1 (45%). At the time of disease progression as assessed by central review, the study blind was broken and all patients were offered the opportunity to take STIVARGA at the investigator’s discretion. Fifty-six (85%) patients randomized to placebo and 41 (31%) patients randomized to STIVARGA received open-label STIVARGA. A statistically significant improvement in PFS was demonstrated among patients treated with STIVARGA compared to placebo (see Table 8 and Figure 2). There was no statistically significant difference in overall survival at the time of the planned interim analysis based on 29% of the total events for the final analysis. Table 8: Efficacy Results for GRID

STIVARGA Placebo (N=133) (N=66) Progression-Free Survival Number of Deaths or Progressions (%) 82 (62%) 63 (96%) Median Progression-Free Survival (months) 4.8 0.9 95% CI (3.9, 5.7) (0.9, 1.1) HR (95% CI) 0.27 (0.19, 0.39) Stratified log-rank test p-value a <0.0001 Overall Survival Number of Deaths (%) 29 (22%) 17 (26%) Median Overall Survival (months) NRb NRb HR (95% CI) 0.77 (0.42, 1.41) Stratified log-rank test p-value a, b 0.2 a Stratified by line of treatment and geographical region. b NR: Not reached.

Reference ID: 4090114

Figure 2: Kaplan-Meier Curves of Progression-Free Survival for GRID

14.3 Hepatocellular Carcinoma (HCC) The clinical efficacy and safety of STIVARGA were evaluated in an international, multicenter, randomized (2:1), double- blind, placebo-controlled trial [Study “REgorafenib after SORafenib in patients with hepatoCEllular carcinoma” (RESORCE); NCT 01774344]. The study enrolled adults with Child-Pugh A and Barcelona Clinic Liver Cancer Stage Category B or C hepatocellular carcinoma, with documented disease progression following sorafenib. The median duration of previous sorafenib treatment was 7.8 months; patients who permanently discontinued sorafenib due to toxicity or were unable to tolerate sorafenib doses of 400 mg once daily were ineligible. Patients were randomized to receive 160 mg regorafenib orally once daily plus best supportive care (BSC) or matching placebo plus BSC for the first 21 days of each 28-day cycle until disease progression or unacceptable toxicity. Randomization was stratified by geographical region (Asia vs rest of world), ECOG performance status (0 vs 1), alpha fetoprotein levels (<400 ng/mL vs ≥400 ng/mL), extrahepatic disease (presence vs absence), and macrovascular invasion (presence vs absence). The major efficacy outcome measure was overall survival (OS). Additional outcome measures were progression-free survival (PFS), overall tumor response rate (ORR) and duration of response as assessed by investigators using RECIST 1.1 and using modified RECIST (mRECIST) for HCC. Patients continued therapy with STIVARGA until clinical or radiological disease progression or unacceptable toxicity. The characteristics of the study population were a median age of 63 years (range 19 to 85 years); 88% male; 41% Asian, 36% White, and 21% not reported; 66% had ECOG performance status (PS) of 0 and 34% had ECOG PS of 1; 98% had Child-Pugh A and 2% had Child-Pugh B. Risk factors for underlying cirrhosis included hepatitis B (38%), alcohol use (25%), hepatitis C (21%), and non-alcoholic steato hepatitis (7%). Macroscopic vascular invasion or extra-hepatic tumor spread was present in 81% of patients. Barcelona Clinic Liver Cancer (BCLC) was stage C in 87% and stage B in 13% of patients. All patients received prior sorafenib and 61% received prior loco-regional transarterial embolization or chemoinfusion procedures. Efficacy results are summarized in Table 9 and Figure 3 below.

Reference ID: 4090114

Table 9: Efficacy Results from Study RESORCE

STIVARGA Placebo n=379 n=194 Overall Survival Number of Deaths (%) 233 (62) 140 (72) Median OS in months (95% CIa) 10.6 (9.1, 12.1) 7.8 (6.3, 8.8) Hazard Ratiob (95% CIa) 0.63 (0.50, 0.79) P-valuec <0.0001 Progression-free Survival (mRECIST) Number of Events (%) 293 (77) 181(93) Progressive Disease 274 (72) 173 (89) Death 19 (5) 8 (4) Median PFS in months (95% CIa) 3.1 (2.8, 4.2) 1.5 (1.4, 1.6) Hazard Ratiob (95% CIa) 0.46 (0.37, 0.56) P-valuec <0.0001 Progression-free Survival (RECIST 1.1) Number of Events (%) 288 (76) 184 (95) Progressive Disease 270 (71) 175 (90) Death 18 (5) 9 (5) Median PFS in months (95% CIa) 3.4 (2.9, 4.2) 1.5 (1.4, 1.5) Hazard Ratiob (95% CIa) 0.43 ( 0.35, 0.52) Overall Response (mRECIST) Overall Response Rate 11% 4% 95% CIa (8%, 14%) (2%, 8%) Complete Response 0.5% 0 Partial Response 10% 4% Overall Response (RECIST 1.1) Overall Response Rate 7% 3% 95% CIa (4%, 10%) (1%, 6%) Complete Response 0 0 Partial Response 7% 3%

a CI=confidence interval. b Estimated with Cox proportional hazard model stratified by geographic region, ECOG performance status, Alpha-fetoprotein level, presence versus absence of extrahepatic disease, and presence versus absence of macrovascular invasion. c Log rank test stratified by geographic region, ECOG performance status, Alpha-fetoprotein level, presence versus absence of extrahepatic disease, and presence versus absence of macrovascular invasion.

Reference ID: 4090114

Figure 3: Kaplan-Meier Curve of Overall Survival from Study RESORCE

16 HOW SUPPLIED/STORAGE AND HANDLING How Supplied Tablets are supplied in packages containing three bottles, with each bottle containing 28 tablets, for a total of 84 tablets per package (NDC 50419-171-03). Storage and Handling Store STIVARGA at 25°C (77°F); excursions are permitted from 15 to 30°C (59 to 86°F) [see USP Controlled Room Temperature]. Store tablets in the original bottle and do not remove the desiccant. Keep the bottle tightly closed after first opening. Discard any unused tablets 7 weeks after opening the bottle. Dispose of unused tablets in accordance with local requirements. 17 PATIENT COUNSELING INFORMATION Advise the patient to read the FDA-approved patient labeling (Patient Information). Hepatotoxicity Advise patients that they will need to undergo monitoring for liver damage and to report immediately any signs or symptoms of severe liver damage to their healthcare provider [see Warnings and Precautions (5.1), Use in Specific Populations (8.6)]. Infections Advise patients to contact their healthcare provider if they experience signs and symptoms of infection [see Warnings and Precautions (5.2)].

Reference ID: 4090114

Hemorrhage Advise patients to contact their healthcare provider for unusual bleeding, bruising, or symptoms of bleeding, such as lightheadedness [see Warnings and Precautions (5.3)]. Gastrointestinal Perforation or Fistula Advise patients to contact a healthcare provider immediately if they experience severe pains in their abdomen, persistent swelling of the abdomen, high fever, chills, nausea, vomiting, or dehydration [see Warnings and Precautions (5.4)]. Dermatologic Toxicity Advise patients to contact their healthcare provider if they experience skin changes including HFSR, rash, pain, blisters, bleeding, or swelling [see Warnings and Precautions (5.5)]. Hypertension Advise patients they will need to undergo blood pressure monitoring and to contact their healthcare provider if blood pressure is elevated or if symptoms from hypertension occur including severe headache, lightheadedness, or neurologic symptoms [see Warnings and Precautions (5.6)]. Cardiac Ischemia and Infarction Advise patients to seek immediate emergency help if they experience chest pain, shortness of breath, feel dizzy, or feel like passing out [see Warnings and Precautions (5.7)]. Reversible Posterior leukoencephalopathy syndrome Advise patients to contact their healthcare provider if they experience signs and symptoms of RPLS [see Warnings and Precautions (5.8)]. Wound Healing Complications Advise patients to contact their healthcare provider if they plan to undergo a surgical procedure or had recent surgery [see Warnings and Precautions (5.9)]. Embryo-Fetal Toxicity Advise patients that regorafenib can cause fetal harm. Advise a pregnant woman of the potential risk to a fetus [see Warnings and Precautions (5.10), Use in Specific Populations (8.1, 8.3)]. Females and Males of Reproductive Potential  Advise women of reproductive potential of the need for effective contraception during STIVARGA treatment and for 2 months after completion of treatment. Instruct women of reproductive potential to immediately contact her healthcare provider if pregnancy is suspected or confirmed during or within 2 months of completing treatment with STIVARGA [see Warnings and Precautions (5.10) and Use in Specific Populations (8.1, 8.3)].  Advise men of reproductive potential of the need for effective contraception during STIVARGA treatment and for 2 months after completion of treatment [see Use in Specific Populations (8.3)]. Lactation Advise nursing mothers that it is not known whether regorafenib is present in breast milk and discuss whether to discontinue nursing or to discontinue regorafenib [see Use in Specific Populations (8.2)]. Administration  Advise patients to swallow the STIVARGA tablet whole with water at the same time each day following a low-fat meal. Inform patients that the low-fat meal should contain less than 600 calories and less than 30% fat [see Dosage and Administration (2.1)].  Advise patients to store medicine in the original container. Do not place medication in daily or weekly pill boxes. Discard any remaining tablets 7 weeks after opening the bottle. Tightly close bottle after each opening and keep the desiccant in the bottle [see How Supplied (16)].

Reference ID: 4090114

Dosing Instructions Advise patients to take STIVARGA after a low fat meal. Advise patients to take any missed dose on the same day, as soon as they remember, and that they must not take two doses on the same day to make up for a dose missed on the previous day [see Dose and Administration (2.1)].

Reference ID: 4090114

Patient Information STIVARGA (sti-VAR-gah) (regorafenib) tablets What is the most important information I should know about STIVARGA? STIVARGA can cause serious side effects, including: Liver problems. STIVARGA can cause liver problems which can be serious and sometimes lead to death. Your healthcare provider will do blood tests to check your liver function before you start taking STIVARGA and during your treatment with STIVARGA to check for liver problems. Tell your healthcare provider right away if you get any of these symptoms of liver problems during treatment:  yellowing of your skin or the white part of your eyes  dark “tea-colored” urine (jaundice)  change in sleep pattern  nausea or vomiting What is STIVARGA? STIVARGA is a prescription medicine used to treat people with:  colon or rectal cancer that has spread to other parts of the body and for which they have received previous treatment with certain chemotherapy medicines  a rare stomach, bowel, or esophagus cancer called GIST (gastrointestinal stromal tumors) that cannot be treated with surgery or that has spread to other parts of the body and for which they have received previous treatment with certain medicines  a type of liver cancer called hepatocellular carcinoma (HCC) in people who have been previously treated with sorafenib It is not known if STIVARGA is safe and effective in children less than 18 years of age. Before taking STIVARGA, tell your healthcare provider about all of your medical conditions, including if you:  have liver problems in addition to liver cancer  have bleeding problems  have high blood pressure  have heart problems or chest pain  plan to have any surgical procedures or have had recent surgery  are pregnant or plan to become pregnant. STIVARGA can harm your unborn baby. o Females should use effective birth control during treatment with STIVARGA and for 2 months after your final dose of STIVARGA. Tell your healthcare provider right away if you become pregnant during treatment with STIVARGA or within 2 months after your final dose of STIVARGA. o Males with female partners who can become pregnant should use effective birth control during treatment with STIVARGA and for 2 months after your final dose of STIVARGA.  are breastfeeding or plan to breastfeed. It is not known if STIVARGA passes into your breast milk. Do not breastfeed during treatment with STIVARGA and for 2 weeks after your final dose of STIVARGA. Tell your healthcare provider about all the medicines you take, including prescription and over-the-counter medicines, vitamins and herbal supplements. STIVARGA may affect the way other medicines work, and other medicines may affect how STIVARGA works. How should I take STIVARGA?  Take STIVARGA exactly as your healthcare provider tells you.  You will usually take STIVARGA 1 time a day for 21 days (3 weeks) and then stop for 7 days (1 week). This is 1 cycle of treatment. Repeat this cycle for as long as your healthcare provider tells you to.  Swallow STIVARGA tablets whole with water following a low-fat meal.  Take STIVARGA at the same time each day following a low-fat meal that contains less than 600 calories and less than 30% fat.  If you miss a dose, take it as soon as you remember on that day. Do not take two doses on the same day to make up for a missed dose.  If you take too much STIVARGA call your healthcare provider or go to the nearest emergency room right away. What should I avoid while taking STIVARGA?  Avoid drinking grapefruit juice and taking St. John’s Wort during treatment with STIVARGA. These can affect the way STIVARGA works.

Reference ID: 4090114

What are the possible side effects of STIVARGA? STIVARGA can cause serious side effects including:  See “What is the most important information I should know about STIVARGA?”  Infection. STIVARGA may lead to a higher risk of infections especially of the urinary tract, nose, throat and lung. STIVARGA may also lead to a higher risk of fungal infections of the mucous membrane, skin or the body. Tell your healthcare provider right away if you get:  fever  burning or pain when urinating  severe cough with or without an increase in mucus  unusual vaginal discharge or irritation (sputum) production  redness, swelling or pain in any part of  severe sore throat the body  shortness of breath  severe bleeding. STIVARGA can cause bleeding which can be serious and sometimes lead to death. Tell your healthcare provider if you have any signs of bleeding during treatment with STIVARGA including:  vomiting blood or if your vomit looks like coffee-grounds  unusual vaginal bleeding  pink or brown urine  nose bleeds that happen often  red or black (looks like tar) stools  bruising  coughing up blood or blood clots  lightheadedness  menstrual bleeding that is heavier than normal  a tear in your stomach or intestinal wall (bowel perforation). STIVARGA may cause a tear in your stomach or intestinal wall (bowel perforation) that can be serious and sometimes lead to death. Tell your healthcare provider right away if you get:  severe pain in your stomach-area (abdomen)  nausea  swelling of the abdomen  vomiting  fever  dehydration  chills  a skin problem called hand-foot skin reaction and severe skin rash. Hand-foot skin reactions are common and sometimes can be severe. Tell your healthcare provider right away if you get redness, pain, blisters, bleeding, or swelling on the palms of your hands or soles of your feet, or a severe rash.  high blood pressure. Your blood pressure should be checked every week for the first 6 weeks of starting STIVARGA. Your blood pressure should be checked regularly and any high blood pressure should be treated during treatment with STIVARGA. Tell your healthcare provider if you have severe headaches, lightheadedness, or changes in your vision.  decreased blood flow to the heart and heart attack. Get emergency help right away if you get symptoms such as chest pain, shortness of breath, feel dizzy or feel like passing out.  a condition called Reversible Posterior Leukoencephalopathy Syndrome (RPLS). Call your healthcare provider right away if you get severe headaches, seizure, confusion, change in vision, or problems thinking.  wound healing problems. If you need to have a surgical procedure, tell your healthcare provider that you are taking STIVARGA. You should stop taking STIVARGA at least 2 weeks before any planned surgery. The most common side effects of STIVARGA include:  pain, including stomach-area (abdomen)  voice changes or hoarseness  tiredness, weakness, fatigue  increase in certain liver function test  frequent or loose bowel movements (diarrhea)  fever  decreased appetite  swelling, pain and redness of the lining in your  infection mouth, throat, stomach and bowel (mucositis)  weight loss Your healthcare provider may change your dose, temporarily stop, or permanently stop treatment with STIVARGA if you have certain side effects. These are not all of the possible side effects of STIVARGA. Call your doctor for medical advice about side effects. You may report side effects to FDA at 1-800-FDA-1088.

Reference ID: 4090114

How do I store STIVARGA?  Store STIVARGA tablets at room temperature between 68° F to 77° F (20° C to 25°C).  Keep STIVARGA in the bottle that it comes in. Do not put STIVARGA tablets in a daily or weekly pill box.  The STIVARGA bottle contains a desiccant to help keep your medicine dry. Keep the desiccant in the bottle.  Keep the bottle of STIVARGA tightly closed.  Safely throw away (discard) any unused STIVARGA tablets after 7 weeks of opening the bottle. Keep STIVARGA and all medicines out of the reach of children. General information about the safe and effective use of STIVARGA. Medicines are sometimes prescribed for purposes other than those listed in a Patient Information leaflet. Do not use STIVARGA for a condition for which it was not prescribed. Do not give STIVARGA to other people even if they have the same symptoms you have. It may harm them. You can ask your healthcare provider or pharmacist for information about STIVARGA that is written for health professionals. What are the ingredients in STIVARGA? Active ingredient: regorafenib Inactive ingredients: cellulose microcrystalline, croscarmellose sodium, magnesium stearate, povidone and colloidal silicon dioxide. Film coat: ferric oxide red, ferric oxide yellow, lecithin (soy), polyethylene glycol 3350, polyvinyl alcohol, talc and titanium dioxide.

Manufactured for Bayer HealthCare Pharmaceuticals Inc., Whippany, NJ 07981 USA. © 2017 Bayer HealthCare Pharmaceuticals Inc. For more information, go to www.STIVARGA-US.com or call 1-888-842-2937. This Patient Information has been approved by the U.S. Food and Drug Administration. Revised: 4/2017

Reference ID: 4090114 CENTER FOR DRUG EVALUATION AND RESEARCH

APPLICATION NUMBER:

203085Orig1s007

SUMMARY REVIEW

Summary Review for Regulatory Action

Date April 28, 2017 From Patricia Keegan Subject Division Director Summary Review NDA Supplement # NDA 203085/S-007 Annlicant Name Baver Healthcare Pha1maceuticals, Inc. Date of Submission October 31, 2016 PDUFA Goal Date April 30, 2017 Proprietary Name I Stivarga tablets/ Established (USAN) Name regorafenib Dosa2e Forms I Stren2th Tablets for oral use/40 mg tablet Proposed Indication(s) STIVARGA is indicated for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with( (b)(4~ Approved Indication STIVARGA is indicated for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with sorafenib Action: Approval

Material Reviewed/Consulted OND Action Package, including: Names of discipline reviewers Regulato1y Project Manager Review Anuja Patel Medical Officer Review Lonaine C. Pelosof Statistical Review Xiaoping (Janet) Jiang Clinical Phannacology Review Vadryn Piene & Y ouwei Bi Nonclinical Phaimacology/Toxicolo!!V Review Whitney Helms OPDP Carole C. Broadnax OSE/DMEPA Otto Townsend Patient Labeling Team Review Morgan A. Walker OND=Office ofNew Drugs OPDP=Office of Prescription Drug Promotion OSE= Office of Surveillance and Epidemiology DMEPA=Division of Medication Error Prevention and Analysis OSI=Office \of Scientific Investigations

Division Director Smnmary Review Page 1of14

Reference ID: 4090125 Division Director Summary Review

1. Introduction

STIVARGA (regorafenib; Bayer Healthcare Pharmaceuticals, Inc.) is a highly promiscuous small molecule inhibitor of multiple membrane-bound and intracellular kinases, including RET, VEGFR1, VEGFR2, VEGFR3, KIT, PDGFR-alpha, PDGFR-beta, FGFR1, FGFR2, TIE2, DDR2, TrkA, Eph2A, RAF-1, BRAF, BRAF V600E, SAPK2, PTK5, Abl and CSF1R. It is approved for: 1) the treatment of patients with metastatic colorectal cancer who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy, and, if RAS wild- type, an anti-EGFR therapy and 2) for the treatment of patients with locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate.

This supplement supports an expansion of the STIVARGA labeling to include a new indication for the treatment of hepatocellular carcinoma in patients with disease progression following sorafenib. The application relies on the results of a single randomized, multicenter, placebo-controlled trial, Study 15982 (RESORCE), to establish the safety and efficacy of regorafenib in this patient population.

Study 15982, entitled: “A randomized, double blind, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib (RESORCE)” was a randomized (1:1), placebo-controlled, international, multicenter trial. Key eligibility criteria were histological or cytological confirmation of HCC or non-invasive diagnosis of HCC per American Association for the Study of Liver Diseases (AASLD) criteria with confirmed cirrhosis; able to tolerate sorafenib; disease progression following sorafenib and randomization within 2 to 10 weeks of the last dose of sorafenib; Barcelona Clinic Liver Cancer (BCLC) stage Category B or C and Child-Pugh class A; not a candidate for resection, local ablation, chemoembolization, or liver transplantation; and no prior systemic therapy other than sorafenib. Patients were randomized (2:1) to receive regorafenib 160 mg po daily or matching placebo administered until disease progression or unacceptable toxicity. Randomization was stratified by geographical region (Asia vs. rest of world (ROW)), ECOG performance status (0 vs. 1), alpha feto-protein level (<400 ng/mL vs. ≥400 ng/mL), extrahepatic disease (presence vs. absence), and macrovascular invasion (presence vs. absence).The primary endpoint was overall survival and secondary endpoints were investigator-assessed progression-free survival and overall response rate using both RECIST v1.1 and modified RECIST for HCC.

A total of 573 patients were randomized to receive regorafenib (n=379) or matching placebo (n=194) across 152 clinical sites in the United States, Australia, Europe and Asia. The demographic characteristics of the study population were a median age of 63 years (range 19 to 85 years); 88% male; 41% Asian, and 36% White, and 21 % not reported. Baseline prognostic factors and tumor characteristics of the study population were 66% had ECOG performance status (PS) of 0 and 34% had ECOG PS of 1; 98% had Child-Pugh A and 2%

Division Director Summary Review Page 2 of 14

Reference ID: 4090125 had Child-Pugh B. Risk factors for underlying cirrhosis included hepatitis B (38%), alcohol use (25%), hepatitis C (21%), and non-alcoholic steatohepatitis (7%). Macroscopic vascular invasion or extra-hepatic tumor spread was present in 81% of patients. Barcelona Clinic Liver Cancer (BCLC) was stage C in 87% and stage B in 13% of patients. All patients received prior sorafenib and 61% received prior loco-regional trans-arterial embolization or trans-arterial chemotherapy infusion procedures.

The trial demonstrated a statistically significant and clinically important improvement in overall survival [hazard ratio (HR) of 0.63 (95% CI: 0.50, 0.79); p<0.0001, stratified log-rank test] with an estimated median survival of 10.8 months in the regorafenib arm and 7.8 months in the placebo arm. Progression-free survival (PFS) was also significantly improved when assessed using mRECIST [HR 0.46 (95% CI: 0.37, 0.56), p<0.0001, stratified log-rank test], with an estimated median PFS of 3.1 months in the regorafenib arm and 1.5 months in the placebo arm. The overall response rate (ORR) was numerically higher for patients randomized to regorafenib (10% vs 4%) compared with the placebo arm. Similar results for PFS and ORR were observed when these endpoints were assessed using RECIST v1.1.

The toxicity of regorafenib in this patient population was similar to that observed in prior approvals and there was no increase in hepatotoxicity among patients with HCC as compared to those with metastatic colorectal cancer or GIST. Among the 374 patients who received at least one dose of regorafenib in the RESORCE trial, 33% were exposed for ≥6 months and 14% were exposed for ≥ 12 months. Dose interruptions for adverse events were required in 58% of regorafenib-treated patients and 48% of patients had their dose reduced. The most common adverse reactions requiring dose modification (interruption or dose reduction) were hand-foot skin reactions (HFSR; also referred to as palmar-plantar erythrodysesthesia syndrome (PPES)) in 20.6%, hyperbilirubinemia in 5.9%, fatigue in 5.1%; and diarrhea in 5.3%. Adverse reactions that resulted in treatment discontinuation were reported in 10.4% of regorafenib-treated patients compared to 3.6% of placebo-treated patients; the most common adverse reactions requiring discontinuation of regorafenib were HFSR/PPES (1.9%) and increased AST (1.6%).

Pooled analyses to better assess the incidence of serious adverse reactions were conducted among the following populations: 1142 regorafenib-treated patients enrolled in one of four placebo-controlled trials; 4518 regorafenib-treated patients across all clinical trials; and 4800 regorafenib-treated patients enrolled in clinical trials or expanded access programs. The most common serious adverse reactions were ≥ Grade 3 hand-foot skin reaction (HFSR) (16%); ≥ Grade 3 infection (9%); and ≥ Grade 3 hemorrhage (3%). The most common adverse drug reactions (≥20%) among 1142 regorafenib-treated patients were pain (including gastrointestinal and abdominal pain), HFSR, asthenia/fatigue, diarrhea, decreased appetite/food intake, hypertension, infection, dysphonia, hyperbilirubinemia, fever, mucositis, weight loss, rash, and nausea.

A substantive issue identified during review of this application was the submission of modified databases based on audits of clinical study sites in China, which raised concerns regarding the reliability of these “cleaned and locked” datasets. These audits resulted in changes to the efficacy results in a small percentage of patients, which did not alter the overall conclusions

Division Director Summary Review Page 3 of 14

Reference ID: 4090125 regardless of the dataset used (original or modified) and sensitivity analyses excluding data from Chinese sites also did not alter the conclusions. Thus, FDA agreed with inclusion of the results of the analyses obtained with the revised datasets in the intent-to-treated (all randomized) population in product labeling. In addition, poor characterization of data variables and description of analysis programs in the application resulted in significant delays in review of the safety information and confirmation of Bayer’s analyses.

2. Background

Indicated Population and Available Therapy The American Cancer Society estimates that there will be 40,710 new cases of hepatocellular and intrahepatic bile duct cancer and 28, 920 deaths due to such cancers in the United States in 2017.1 The 5-year survival rate for patients with regional involvement is 11% and for those with metastatic disease is 3%.2

Sorafenib is the only systemic drug that is FDA-approved the treatment of hepatocellular cancer. Sorafenib was approved for “the treatment of patients with unresectable hepatocellular carcinoma (HCC)” on November 16, 2007. This approval was based on demonstration of improved survival [HR 0.69 (95% CI: 0.55, 0.87); p=0.00058] in an international, multicenter, randomized (1:1), double blind, placebo-controlled trial in 602 patients with unresectable hepatocellular carcinoma. The median survival was 10.7 months in the sorafenib arm and 7.9 months in the placebo arm. The trial also demonstrated an improvement in PFS [HR 0.58 (95% CI: 0.45, 0.74); p<0.0001] with median PFS of 5.5 months and 2.8 months in the sorafenib and placebo arms, respectively.

Regulatory History The clinical development program for the proposed indication was conducted under IND 75642, submitted to FDA in June 2006.

On October 3, 2012, Bayer submitted a new clinical protocol for Study 15982, entitled: “A randomized, double blind, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib (RESORCE)” to IND 75642. The design of this protocol was not discussed in an end-of-Phase 2 meeting.

On December 14, 2012, FDA issued an Advice/Information Request letter, containing comments and requests for additional information regarding the RESORCE study. FDA advised that in a future marketing application, Bayer should provide justification that the results of this trial can be extrapolated to the U.S. population and to revise the analysis plan to provide an adjustment for multiplicity if claims would be sought for secondary efficacy endpoints.

1 https://www.cancer.org/cancer/liver-cancer/about/what-is-key-statistics html 2 https://cancerstatisticscenter.cancer.org/#/cancer-site/Liver%20and%20intrahepatic%20bile%20duct

Division Director Summary Review Page 4 of 14

Reference ID: 4090125 On July 26, 2013, Bayer submitted an amended protocol for the RESORCE trial address FDA's comments in the December 14, 2012, letter. The protocol was further amended and submitted to IND 75642 on Febrnaiy 16, 2016, and on December 18, 2015.

On December 18, 2015, Bayer submitted a statistical analysis plan (SAP) for the RESORCE protocol and a revised clinical protocol.

On Mai·ch 3, 2016, FDA issued Written Responses to a Type C meeting request, providing preliminaiy advice on the content and fonnat of the planned efficacy supplement to be based on the results of Study 15982. The results of the ti·ial were not included in the meeting briefing package (final analysis projected to occur in Q2 2016)

On April 21 , 2016, Bayer submitted a US-specific SAP, dated April 20, 2016, to address FDA comments in the December 14, 2012, Advice letter. Bayer stated that the database lock was planned for April 22, 2016.

On July 21, 2016, a pre-sNDA meeting was held to discuss the results of Study 15982 and the proposed contents of the planned supplement. FDA agreed that the results of Study 15982, suppo1ted by activity in a single aim h'ial (Study 14596) provided sufficient info1m ation to allow FDA to evaluate the efficacy and safety of regorafenib in HCC. However, FDA did not agree with the p~posed indication and stated! (b)C-0 ______....,.. ______..,.... ______,,,____ ...,...... ------but agreed to consider alternate wording supported by "real world evidence" during review of the supplement. FDA agreed with the proposed approach to provide assessment of exposure response analyses and assessment in subgroups (Asian vs. non-Asian). Finally, FDA agreed on the strategy for a planned rolling review based on possible Fast Track designation.

On July 28, 2016, FDA granted Fast Track Designation for the clinical development program investigating regorafenib for the "treatment of patients with (bT<' hepatocellular cai·cinoma (HCC) who have been previously treated with sorafenib."

On September 1, 2016 submission, Bayer requested pe1mission to submit po1tions of the application (rolling review). In this request, Bayer provided a schedule for submission. FDA granted this request on October 4, 2016.

Division Director Smnmary Review Page 5of14

Reference ID: 4090125 On October 6, 2016, Bayer submitted information to Modules 1 and 5 and completed the supplement with a submission on October 30, 2016, and was given priority review designation on December 29, 2016.

3. CMC/ Biopharmaceutics/Device

Not applicable.

4. Nonclinical Pharmacology/Toxicology

I concur with the conclusions reached by the non-clinical pharmacology/toxicology reviewer that there are no outstanding issues that preclude approval. Bayer provided the results of additional non-clinical studies to support modifications to section 12.1 of the US package insert. These included data to support inclusion of CSF1R as a target of regorafenib that is clinically relevant, data to support effects on tumor immunity as a potential mechanism of action, and the results of pharmacology studies characterizing anti-tumor activity.

5. Clinical Pharmacology/Pharmacogenomics

I concur with the conclusions reached by the clinical pharmacology/biopharmaceutics reviewer that there are no outstanding clinical pharmacology issues that preclude approval. Bayer conducted and provided the results of a post-hoc population pharmacokinetic analysis. The reviewers concluded that this analysis did not identify clinically important differences in the mean total exposure of regorafenib (i.e., regorafenib plus its active metabolites, M2 and M5) in patients with normal hepatic function or with mild or moderate hepatic impairment. There were no patients with severe hepatic impairment (total bilirubin >3x ULN or Child-Pugh C) enrolled in the RESORCE trial, thus there remains no information regarding the pharmacokinetics of regorafenib in patients with severe hepatic impairment. In exposure- response analyses, no relevant relationships were identified between exposure and either efficacy or toxicity.

6. Clinical Microbiology

Not applicable.

Division Director Summary Review Page 6 of 14

Reference ID: 4090125 7. Clinical/Statistical-Efficacy

This supplement was supported by a single, large, multicenter, placebo-controlled efficacy trial for a supplemental indication. Bioresearch monitoring of clinical study sites was not requested since the trial used a primary efficacy endpoint that is objective and not subject to bias (overall survival) and was statistically robust and supported by treatment effects on secondary efficacy endpoints. Additionally, this is the third trial with regorafenib to demonstrate improvement in survival in patients with cancer (the other settings were treatment of gastrointestinal stromal tumors and metastatic colorectal cancer).

Study Design This application was supported by the results of a single clinical trial, Study 15982 (RESORCE), titled “A randomized, double blind, placebo-controlled, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib.”

The key inclusion criterion were either histological or cytological confirmation of HCC or non-invasive diagnosis of HCC per American Association for the Study of Liver Diseases (AASLD) criteria with confirmed cirrhosis; able to tolerate sorafenib at 400 mg daily for at least 20 days with disease progression following sorafenib and randomization within 2 to 10 weeks of the last dose of sorafenib; Barcelona Clinic Liver Cancer (BCLC) stage Category B or C and Child-Pugh class A; not a candidate for resection, local ablation, chemoembolization, or liver transplantation; no prior systemic therapy other than sorafenib. Patients with any of the following were ineligible: bleeding from esophageal varices, uncontrolled ascites or pleural effusions, uncontrolled hypertension; ≥ Grade 3 bleeding within 30 days prior to randomization; arterial or venous thromboembolic events within 6 months prior to randomization; ≥ Grade 3 proteinuria; non-healing wound or bone fracture; hepatitis B infection or hepatitis A infection requiring anti-viral therapy; symptomatic interstitial lung disease; or HIV infection.

Patients were randomized (2:1) to receive  regorafenib 160 mg orally, once daily for days 1-21 of each 28-day cycle until disease progression, unacceptable toxicity OR  Matching placebo orally, once daily for days 1-21 of each 28-day cycle until disease progression, unacceptable toxicity

Randomization was stratified by geographical region (Asia vs. rest of world (ROW)), ECOG performance status (0 vs. 1), alpha feto-protein level (<400 ng/mL vs. ≥400 ng/mL), extrahepatic disease (presence vs. absence), and macrovascular invasion (presence vs. absence).

Patients were assessed for disease state by CT or MRI every 6 weeks for the first 8 cycles then every 12 weeks thereafter. Assessment of tumor-based endpoints was conducted by

Division Director Summary Review Page 7 of 14

Reference ID: 4090125 investigators and by an independent radiologic review using RECIST v1.1 and modified RECIST for HCC.3

The primary endpoint was overall survival. The assumptions for the study sample size of 560 patients were a true hazard ratio of 0.70 for overall survival, median survival of 8 months in the placebo arm and 11.4 months in the regorafenib arm, and a requirement for 370 deaths to provide 90% power to detect statistically significant difference in survival at a 2-sided significance level of 0.05. Secondary efficacy endpoints were time-to-progression (TTP), progression-free survival (PFS), and overall response rate (ORR).

Results A total of 573 patients were randomized to receive regorafenib (n=379) or matching placebo (n=194) across 152 clinical sites in the United States, Australia, Europe and Asia. The first visit of the first patient was on May 14, 2013 and the last visit of the last patient was on February 29, 2016. The “study completion date” was February 29, 2016.

The demographic characteristics of the study population were a median age of 63 years (range 19 to 85 years); 88% male; 41% Asian, and 36% White, and 21 % not reported. Baseline prognostic factors and tumor characteristics of the study population were 66% had ECOG performance status (PS) of 0 and 34% had ECOG PS of 1; 98% had Child-Pugh A and 2% had Child-Pugh B. Risk factors for underlying cirrhosis included hepatitis B (38%), alcohol use (25%), hepatitis C (21%), and non-alcoholic steatohepatitis (7%). Macroscopic vascular invasion or extra-hepatic tumor spread was present in 81% of patients. Barcelona Clinic Liver Cancer (BCLC) was stage C in 87% and stage B in 13% of patients. All patients received prior sorafenib and 61% received prior loco-regional trans-arterial embolization or trans-arterial chemotherapy infusion procedures. The following table, abstracted from product labeling, summarizes the key efficacy results The results for tumor-based endpoints as assessed by investigators according to mRECIST and RECIST v1.1 were similar.

3 Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepatocellular carcinoma. Semin Liver Dis 2010;30:52-60.

Division Director Summary Review Page 8 of 14

Reference ID: 4090125 (b)(4) Efficacy Results H RESORCE I (6)\4~ Placebo n=379 n=194 Overall Survival Number of Deaths(%) 233 (62) 140 (72) Median OS in months (95% CI3) 10.6 (9. l , 12.1) 7.8 (6.3, 8.8) Hazard Ratiob (95% era) 0.63 (0.50, 0.79) p-valuec <0.0001 Proe:ression-free Survival (mRECrST) Number of Events(%) 293 (77) 181(93) Progressive Disease 274 (72) 173 (89) Death 19 (5) 8 (4) Median PFS in months (95% e ra) 3. 1 (2.8, 4.2) 1.5 (1.4, 1.6) Hazard Ratiob (95% CI3) 0.46 (0.37, 0.56) p-valuec <0.0001 Progression-free Survival (RECrST I.I) Number of Events (%) 288 (76) 184 (95) Progressive Disease 270 (71) 175 (90) Death 18 (5) 9 (5) Median PFS in months (95% CI3) 3.4 (2.9. 4.2) 1.5 (1.4. 1.5) Hazard Ratiob (95% e ra) 0.43 ( 0.35, 0.52) Overall Response (mRECrsn Overall Response Rate 11% 4% 95% CI3 (8%, 14%) (2%, 8%) Complete Resoonse 0.5% 0 Partial Response 10% 4% Overall Response (RECrST 1.1) Overall Response Rate 7% 3% 95% CI3 (4%, 10%) (1%, 6%) Complete Response 0 0 Partial Response 7% 3% CI=confidence interval. b Estimated with Cox proportional hazard model stratified by geographic region, ECOG performance status, Alpha-fetoprotein level, presence versus absence of extrahepatic disease, and presence versus absence of macrovascular invasion. c Log rank test stratified by geographic region, ECOG perfo1mance status, alpha-fetoprotein level, presence versus absence of extrahepatic disease, and presence versus absence of macrovascular mvas1on.

Division Director Smnmary Review Page 9of14

Reference ID: 4090125 Kaplan-Meier Curv CbT RESORCE

100 -- STIVARGA (n: 379) ---- · Placebo (n=194)

75 .. _ l .?;- ' ' '• ~ ·...... Q"' , 2 50 Q. ·. '1V ' ' ·:;:> !:i en ' '"- .. -._ ' - 25 ... ,_ 1 1 .._ _ _ _ -,_ --, '• , .. ______,

0 3 6 9 12 15 18 21 24 27 30 33 Months From Randomization Patients at risk STIVARGA 31 6 224 170 122 78 54 34 21 10 4 Placebo 149 95 62 37 26 16 8 5 3

A substantive data. issue that occmTed during review of this application was the submission of modified databases based on audits of clinical study sites in China, which raised concerns regarding the reliability of these "cleaned and locked" datasets. These audits resulted in changes to the efficacy results in a small percentage of patients. The statistical reviewer conducted a sensitivity analysis for overall survival that excluded the 137 patients enrolled across 27 clinical study sites in China. The results of this sensitivity analysis are summarized in the table below.

Sensitivity Analysis Excluding Patients Enrolled in China

Regorafenib Placebo (n=291) (n=145) Number of Event (%) 183 108 Number of Censored (%) 108 37 Median OS in months (95% CI) 10.9 (9 .1, 13.2) 8.3 (6.8.9.3) Hazard ratio (95%CI) 0.62 (0.48, 0.80)

Division Director Smnmary Review Page 10of14

Reference ID: 4090125 Based on comparisons of the results in the original dataset, the revised dataset, and the sensitivity analysis (above), the results were not substantially altered with the revised dataset. Based on Bayer’s statements that the revised dataset is a more accurate reflection of the data and analyses with this revised dataset resulted in a slightly smaller treatment effect (based on the hazard ratio) than that provided in the original submission, the clinical and statistical review staff agreed that the results presented in the product labeling should be based on the revised datasets.

8. Safety

Size of the database The size of the safety database (374 regorafenib-treated patients) was sufficient to make a risk: benefit assessment in the indicated population of patients receiving second-line treatment for unresectable or metastatic hepatocellular carcinoma). The toxicity of regorafenib in this patient population was similar to that observed in prior approvals and there was no increase in hepatotoxicity among patients with HCC as compared to those with metastatic colorectal cancer or GIST.

Among the 374 patients who received at least one dose of regorafenib in the RESORCE trial, 33% were exposed for ≥6 months and 14% were exposed for ≥ 12 months. Dose interruptions for adverse events were required in 58% of regorafenib-treated patients and 48% of patients had their dose reduced. The most common adverse reactions requiring dose modification (interruption or dose reduction) were hand-foot skin reactions (HFSR; also referred to as palmar-plantar erythrodysesthesia syndrome (PPES)) in 20.6%, hyperbilirubinemia in 5.9%, fatigue in 5.1%; and diarrhea in 5.3%. Adverse reactions that resulted in treatment discontinuation were reported in 10.4% of regorafenib-treated patients compared to 3.6% of placebo-treated patients; the most common adverse reactions requiring discontinuation of regorafenib were HFSR/PPES (1.9%) and increased AST (1.6%).

Major safety concerns related to labeling The toxicity of regorafenib in this setting was similar to that observed in prior approvals. Pooled analyses to better assess the incidence of serious adverse reactions were conducted among the following populations: 1142 regorafenib-treated patients enrolled in one of four placebo-controlled trials; 4518 regorafenib-treated patients across all clinical trials; and 4800 regorafenib-treated patients enrolled in clinical trials or expanded access programs. These included hepatotoxicity (ranging from 0 to 1.6% incidence); ≥ Grade 3 infection (9%); ≥ Grade 3 hemorrhage (3%); gastrointestinal perforation (0.6%); gastrointestinal fistula (0.8%); ≥Grade 3 hand-foot skin reaction (HFSR) (16%); ≥Grade 3 rash (3%); Stevens Johnson syndrome and erythema multiforme (<0.1% each) and toxic epidermal necrolysis (<0.02%); hypertensive crisis (0.2%); myocardial ischemia and infarction (0.9%); r reversible posterior leukoencephalopathy syndrome (RPLS) (0.02%). In addition, the regorafenib can cause embryo-fetal toxicity and, based on effects in the class of vascular endothelial growth factor inhibitors, is expected to impair wound healing.

REMS

Division Director Summary Review Page 11 of 14

Reference ID: 4090125 I concur with the clinical review team that no new safety issues were identified and that the risks of regorafenib in this patient population are acceptable in light of the life-threatening stage of disease and lack of alternative therapy. I concur that REMS (risk mitigation and evaluation strategies) are not required to ensure safe use of this marketed drug in the indication population (patients with HCC with disease progression following sorafenib).

PMRs and PMCs The review team did not identify the need for post-marketing requirements under 505(o) and there were no post-marketing commitments requested by FDA for this development program.

9. Advisory Committee Meeting

This efficacy supplement was not referred for review to the Oncologic Drugs Advisory Committee because this is not the first approval for this drug, the safety profile is acceptable for this indication, the clinical trial design is similar to other products approved for this indication, and outside expertise was not necessary since there were no controversial issues that would benefit from advisory committee discussion. . 10. Pediatrics

On June 4, 2015, FDA designed regorafenib as an orphan drug for the “treatment of hepatocellular carcinoma.” Based on this action, this efficacy supplement is exempt for the requirements of the Pediatric Research Equity Act (PREA) for the proposed indication.

11. Other Relevant Regulatory Issues

There are no other unresolved relevant regulatory issues.

12. Labeling

 Physician labeling o Across all labeling sections, the references to clinical studies now identify the study acronym (and in section 14, refer to the NCT numeric designation at www.clinicaltrials.gov). This change is based on feedback from external users who had difficulty determining the study being cited in the USPI. o Indications and Usage: The proposed indication (treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4) was revised to more accurately reflect the patient population studied (previously treated with sorafenib).

Division Director Summary Review Page 12 of 14

Reference ID: 4090125 o Dosage and Administration: modified to clarify the directions for dosing after/following a low fat meal (similar edits to Patient Counseling and Patient Package Insert); modified for consistency and clarity regarding dose modifications in patients with infections. o Warnings and Precautions: The Warnings and Precautions section was updated by Bayer to include a new subsection on Infections, identified based on the pooled analysis of adverse reactions among 1142 regorafenib-treated patients enrolled in randomized, placebo-controlled studies. The incidence of adverse reactions, including ≥ Grade 3 adverse reactions was updated to reflect the incidence across the pooled analysis (with the exception of hepatotoxicity) to provide a more precise estimation of these uncommon serious adverse reactions. The risks of hepatotoxicity were provided by indication rather than as a pooled analysis, since such risks may differ based on the extent of underlying pathology in the liver and that this would be of concern to prescribers. It is noted, however, that the incidence of hepatotoxicity was actually not substantially different across the indicated patient populations. Finally, subsection on Gastrointestinal Perforation and Fistula moved to increase prominence given the severity (fatal events) of these adverse reactions. o Adverse Reactions: This section was updated to include the characterization of the safety population (as treated) and adverse reactions and clinical laboratory abnormalities observed in the RESORCE trial. Change to the incidence of “pain” in the GRID and CORRECT trials based on re-analysis using multiple preferred terms in this composite endpoint and for consistency with results reported for RESORCE. o Use in Specific Populations: Geriatrics subsection updated to reflect data in the pooled analysis of 1142 regorafenib-treated patients and to describe the apparent increased in risk of Grades 3-4 hypertension in elderly patients as compared to younger patients (relative to placebo-treated patients) observed in the RESORCE trial. Edits to Renal Impairment subsection (b) (4) and revisions to subsection on Hepatic Impairment for clarity and consistency with data reviewed in the updated population PK analyses. o Clinical Pharmacology: Section 12.1 was updated to include CSF1R as a target of regorafenib, tumor immunity as a potential mechanism of action, and the results of pharmacology studies characterizing anti-tumor activity. Section 12.3 was updated to describe the pharmacokinetics of regorafenib and its major active metabolite in patients with hepatic impairment. o Nonclinical Pharmacology/Toxicology o Clinical Studies: Updated to include the results of the RESORCE trial. Results were expanded to provide additional details on the clinical trial design and characteristics for the study population. The results for tumor-based endpoints based on both mRECIST for HCC and RECIST v1.1 as prescribers may be interested in the differences (if any) according to the response criteria used, however based on the FDA’s interpretation of the protocol and analysis plan (which were not explicit) the primary analysis of tumor- based endpoint was to be based on mRECIST and statistical tests were limited to PFS were limited to this analysis (there was no adjustment for testing of ORR). o Patient Counseling: Edited for conformance with current labeling practices and applicable guidances and to reflect re-ordering of the Warnings and Precautions section of product labeling.

Division Director Summary Review Page 13 of 14

Reference ID: 4090125  Patient labeling/Medication guide: Patient labeling was revised to reflect the new indication and description of adverse reactions observed in the RESORCE trial, as reflected in the full prescribing information. .

13. Decision/Action/Risk Benefit Assessment

 Regulatory Action: Approval

 Risk Benefit Assessment Hepatocellular carcinoma is a serious and life-threatening disease with a 5-year survival rate of 3% in patients with metastatic disease, not amenable to local therapy. There is only one drug (sorafenib) that is FDA-approved for this population; the RESORCE trial enrolled patients who were no longer responding to sorafenib. Thus, this population has a clear, unmet medical need.

The RESORCE trial demonstrated a statistically robust and clinically important improvement in overall survival, as characterized by an increase in median survival of approximately 3 months (10.6 months compared with 7.8 months) supported by a statistically robust doubling in progression-free survival (3.1 months compared with 1.5 months). The toxicity profile of regorafenib was similar to that observed in the previously approved indications, with the most common adverse reactions resulting in treatment modification being HFSR, fatigue, diarrhea, and hepatic laboratory abnormalities. While regorafenib can results in serious and fatal adverse reactions, the improvement in survival indicates that the benefits of tumor control outweighs these serious risks, which are considered acceptable by the patient and medical oncology community in light of the serious and fatal nature of metastatic/ unresectable HCC.

 Recommendation for Postmarketing Risk Evaluation and Mitigation Strategies There were no new safety signals identified during review of this efficacy supplement that required Risk Evaluation and Mitigation Strategies under 505(o) to ensure safe and effective use.

 Recommendation for other Postmarketing Requirements and Commitments I concur that there are no safety signals that require post-marketing assessment and no post-marketing commitments were requested by FDA based on the information provided in this supplement.

Division Director Summary Review Page 14 of 14

Reference ID: 4090125 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------PATRICIA KEEGAN 04/27/2017

Reference ID: 4090125 CENTER FOR DRUG EVALUATION AND RESEARCH

APPLICATION NUMBER:

203085Orig1s007

OFFICER/EMPLOYEE LIST Officer/Employee List Application: sNDA 203085/ Supplement 007

The following officers or employees of the FDA participated in the decision to approve this application and consented to be identified on this list:

Patricia Keegan Steven Lemery Lorraine Pelosof Anwar Goheer Whitney Helms Xiaoping (Janet) Jiang Lisa Rodriguez Rajeshwari Sridhara Vadryn Pierre Jeanne Fourie Zirkelbach Hong Zhao Youwei Bi Jiang Liu Lorenzo Rocco Ramesh Raghavachari Hasmukh B. Patel Otto Townsend Chi-Ming (Alice) Tu Carole Broadnax Barbara Fuller Morgan Walker Latonia Ford Anuja Patel Monica Hughes Brantley Dorch

Reference ID: 4092559 CENTER FOR DRUG EVALUATION AND RESEARCH

APPLICATION NUMBER:

203085Orig1s007

MEDICAL REVIEW(S) CLINICAL REVIEW- ADDENDUM

Application Type sNDA Application Number(s) 203085/155, 156 Priority or Standard Priority

Submit Date(s) October 6, 2016 October 30, 2016 Received Date(s) October 6, 2016 October 31, 2016 PDUFA Goal Date April 28, 2017 Division / Office DOP2/OHOP

Reviewer Name(s) Lorraine Pelosof Review Completion Date April 28, 2017

Established Name Regorafenib (Proposed) Trade Name Stivarga Therapeutic Class small molecule kinase inhibitor Applicant Bayer HealthCare Pharmaceuticals Inc.

Formulation(s) 40 mg tablets Dosing Regimen 4 x 40 mg tablets every day; 3 weeks on/1 week off in each 4 week cycle Indication(s) for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4) Intended Population(s) patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4)

Reference ID: 4091692 This review addendum addresses and updates various review issues that arose after the completion of the Clinical Review document dated 4/711 7.

1 Recommendations/Risk Benefit Assessment

1.1 Recommendation on Regulatory Action

After fmi her discussions and review, the indication was revised to:

Approval is recommended, pending agreement on final labeling, for the use of regorafenib for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously ti·eated with sorafenib.

This decision was based on the fact that a •rc-0 did not reflect the data submitted and could lead to CbTC'

6 Review of Efficacy

6.1.4 Analysis of Primary Endpoint

On April 20, 2017, Bayer info1m ed FDA about an internal audit of some of their clinical sites in China that resulted in Bayer updating survival data. According to Bayer, inclusion of the updated data would have resulted in a result slightly more favorable to regorafenib. A different sensitivity analysis was perfo1m ed by the Office of Biostatistics (removing data from clinical sites that em olled patients in China) that did not reveal a significant impact due to this update (hazard ratio for overall survival decreased from 0.63 to 0.62 favoring the regorafenib aim ). Given the magnitude and direction of the difference, the label will reflect the analysis based on the original data. .

7 Review of Safety

On page 39 of the clinical review, 3£d pai·agraph, the text should state "On March 23, 2017" instead of "May 23, 2017".

7.3.3 Dropouts and/or Discontinuations

After fmi her clai·ifications regai·ding the data included within the datasets, FDA agreed with Bayer's incidence numbers for regorafenib dose inte1111ptions (58%), reductions (48%), and discontinuations (10%) due to adverse events.

9 Appendices

9.1 Labeling Recommendations

The following summai·izes key labeling updates made since the time of submission of the Clinical Review on 4/711 7.

Reference ID: 4091692 1 INDICATIONS AND USAGE 1.3 Hepatocellular Carcinoma

As indicated above, the indication agreed upon is “for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with sorafenib”.

5 WARNINGS AND PRECAUTIONS

The following populations were defined in the label:

4,518 = The number of patients treated with single- agent regorafenib across all the clinical trials.

4,800 = The number of patients treated with regorafenib across all the clinic trials or in an expanded access program.

6 ADVERSE REACTIONS 6.1 Clinical Trials Experience

The list of the most common adverse drug reactions (≥20%) in patients receiving regorafenib in randomized, placebo-controlled trials was confirmed as: pain (including gastrointestinal and abdominal pain), HFSR, asthenia/fatigue, diarrhea, decreased appetite/food intake, hypertension, infection, dysphonia, hyperbilirubinemia, fever, mucositis, weight loss, rash, and nausea.

8 USE IN SPECIFIC POPULATIONS 8.5 Geriatric Use

Regarding the incidence of hypertension in patients ≥ 65 years of age who received regorafenib, Bayer agreed to examine this issue further in the pooled population from its randomized, placebo-controlled trials and the following was added to the label: “There was an increased incidence of Grade 3 hypertension (18% versus 9%) in the placebo- controlled trials among STIVARGA-treated patients 65 years of age and older as compared to younger patients. In addition, one Grade 4 hypertension event has been reported in the 65 years and older age group and none in the younger age group.”

14 CLINICAL STUDIES 14.3 Hepatocellular Carcinoma (HCC)

Section 14 of the label was updated to reflect data from the RESORCE trial.

Reference ID: 4091692 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------LORRAINE C PELOSOF 05/01/2017

STEVEN J LEMERY 05/01/2017

Reference ID: 4091692 CLINICAL REVIEW

Application Type sNDA Application Number(s) 203085/155, 156 Priority or Standard Priority

Submit Date(s) October 6, 2016 October 30, 2016 Received Date(s) October 6, 2016 October 31, 2016 PDUFA Goal Date April 28, 2017 Division / Office DOP2/OHOP

Reviewer Name(s) Lorraine Pelosof Review Completion Date April 6, 2017

Established Name Regorafenib (Proposed) Trade Name Stivarga Therapeutic Class small molecule kinase inhibitor Applicant Bayer HealthCare Pharmaceuticals Inc.

Formulation(s) 40 mg tablets Dosing Regimen 4 x 40 mg tablets every day; 3 weeks on/1 week off in each 4 week cycle Indication(s) for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated (b) (4) with Intended Population(s) patients with hepatocellular carcinoma (HCC) who have been (b) (4) previously treated with

Reference ID: 4081175 APPEARS THIS WAY ON ORIGINAL

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table of Contents

1 RECOMMENDATIONS/RISK BENEFIT ASSESSMENT ...... 8 1.1 Recommendation on Regulatory Action ...... 8 1.2 Risk Benefit Assessment ...... 8 1.3 Recommendations for Postmarket Risk Evaluation and Mitigation Strategies . 12 1.4 Recommendations for Postmarket Requirements and Commitments ...... 12 2 INTRODUCTION AND REGULATORY BACKGROUND ...... 12 2.1 Product Information ...... 12 2.2 Currently Available Treatments for Proposed Indications ...... 13 2.3 Availability of Proposed Active Ingredient in the United States ...... 13 2.4 Important Safety Issues With Consideration to Related Drugs ...... 13 2.5 Summary of Presubmission Regulatory Activity Related to Submission ...... 14 2.6 Other Relevant Background Information ...... 14 3 ETHICS AND GOOD CLINICAL PRACTICES ...... 16 3.1 Submission Quality and Integrity ...... 16 3.2 Compliance with Good Clinical Practices ...... 16 3.3 Financial Disclosures ...... 16 4 SIGNIFICANT EFFICACY/SAFETY ISSUES RELATED TO OTHER REVIEW DISCIPLINES ...... 17 4.1 Chemistry Manufacturing and Controls ...... 17 4.2 Clinical Microbiology ...... 17 4.3 Preclinical Pharmacology/Toxicology ...... 17 4.4 Clinical Pharmacology ...... 17 4.4.1 Mechanism of Action ...... 18 4.4.2 Pharmacokinetics ...... 18 5 SOURCES OF CLINICAL DATA...... 18 5.1 Tables of Studies/Clinical Trials ...... 18 5.2 Review Strategy ...... 18 5.3 Discussion of Study 15982 ...... 19 6 REVIEW OF EFFICACY ...... 28 Efficacy Summary ...... 28 6.1 Indication ...... 29 6.1.1 Methods ...... 29 6.1.2 Demographics ...... 30 6.1.3 Subject Disposition ...... 33 6.1.3.1 Protocol violations and deviations ...... 33 6.1.4 Analysis of Primary Endpoint ...... 33 6.1.5 Analysis of Secondary Endpoints ...... 34

3 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

6.1.6 Other Endpoints ...... 35 6.1.7 Subpopulations ...... 37 6.1.8 Analysis of Clinical Information Relevant to Dosing Recommendations .... 38 6.1.9 Discussion of Persistence of Efficacy and/or Tolerance Effects ...... 38 6.1.10 Additional Efficacy Issues/Analyses ...... 38 7 REVIEW OF SAFETY ...... 38 Safety Summary ...... 38 7.1 Methods ...... 39 7.1.1 Studies/Clinical Trials Used to Evaluate Safety ...... 39 7.1.2 Categorization of Adverse Events ...... 39 7.1.3 Pooling of Data Across Studies/Clinical Trials to Estimate and Compare Incidence ...... 40 7.2 Adequacy of Safety Assessments ...... 40 7.2.1 Overall Exposure at Appropriate Doses/Durations and Demographics of Target Populations ...... 40 7.2.2 Explorations for Dose Response ...... 40 7.2.3 Special Animal and/or In Vitro Testing ...... 40 7.2.4 Routine Clinical Testing ...... 40 7.2.5 Metabolic, Clearance, and Interaction Workup ...... 40 7.2.6 Evaluation for Potential Adverse Events for Similar Drugs in Drug Class .. 40 7.3 Major Safety Results ...... 41 7.3.1 Deaths ...... 45 7.3.2 Nonfatal Serious Adverse Events ...... 47 7.3.3 Dropouts and/or Discontinuations ...... 49 7.3.4 Significant Adverse Events- Non-Fatal Grade 3-4 AEs ...... 50 7.3.5 Submission Specific Primary Safety Concerns ...... 56 7.4 Supportive Safety Results ...... 57 7.4.1 Common Adverse Events ...... 57 7.4.2 Laboratory Findings ...... 58 7.4.3 Vital Signs ...... 59 7.4.4 Electrocardiograms (ECGs) ...... 59 7.4.5 Special Safety Studies/Clinical Trials ...... 59 7.4.6 Immunogenicity ...... 60 7.5 Other Safety Explorations ...... 60 7.5.1 Dose Dependency for Adverse Events ...... 60 7.5.2 Time Dependency for Adverse Events ...... 60 7.5.3 Drug-Demographic Interactions ...... 60 8 POSTMARKET EXPERIENCE ...... 62 9 APPENDICES ...... 62 9.1 Labeling Recommendations ...... 62 9.2 Advisory Committee Meeting ...... 64 9.3 Additional Tables ...... 64

4 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

APPEARS THIS WAY ON ORIGINAL

5 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table of Tables

Table 1- Study 15982: Protocol Amendments ...... 27 Table 2- Study 15982: Demographics (ITT population) ...... 30 Table 3-Study 15982: Disease characteristics (ITT population) ...... 31 Table 4- Study 15982: Overall survival results (ITT) ...... 34 Table 5- Study 15982: PFS results (ITT population) ...... 35 Table 6- Study 15982: ORR results (ITT population) ...... 35 Table 7-Study 15982: Major safety results summary ...... 41 Table 8- Study 15982: AEs by SOC ...... 42 Table 9- Study 15982: AEs by PT with ≥ 5% incidence in regorafenib arm and with a ≥ 3% difference between arms ...... 43 Table 10- Study 15982: AEs by HLT (incidence ≥ 5%) ...... 44 Table 11- Study 15982: Causes of Grade 5 events by PT ...... 46 Table 12- Causes of Grade 5 events by SOC ...... 47 Table 13- Study 15982: Non-fatal (Grades 1-4) SAEs (incidence ≥1%) by PT ...... 48 Table 14- Study 15982: Grade 3-4 SAEs by SOC ...... 48 Table 15- Study 15982: Study drug dose reduced, interrupted, or withdrawn ...... 50 Table 16- Study 15982: Grade 3-4 AEs by SOC ...... 50 Table 17- Study 15982: Grade 3-4 AEs by PT (incidence ≥ 2%) ...... 51 Table 18- Study 15982: Grade 3-4 Investigations AEs by PT (and incidence ≥1% in regorafenib arm) ...... 52 Table 19- Study 15982: Grade 3-4 Metabolism and Nutrition Disorder AEs by PT (and incidence ≥1% in regorafenib arm) ...... 53 Table 20- Study 15982: Grade 3-4 Gastrointestinal Disorder AEs by PT (and incidence ≥1% in regorafenib arm) ...... 54 Table 21- Study 15982: Grade 3-4 Vascular Disorder AEs by PT (and incidence ≥1% in regorafenib arm) ...... 54 Table 22- Study 15982: Grade 3-4 Skin and subcutaneous tissue disorder AEs by PT (and incidence ≥1% in regorafenib arm) ...... 55 Table 23- Study 15982: Grade 3-4 General disorders and administration site condition AEs by PT (and incidence ≥1% in regorafenib arm) ...... 55 Table 24- Study 15982: Grade 3-4 Hepatobiliary disorder AEs by PT ...... 55 Table 25- Study 15982: Grade 3-4 Infections and infestation AEs by PT (and incidence ≥1% in regorafenib arm) ...... 56 Table 26- Study 15982: Grade 1-2 AEs by PT (incidence > 5%) ...... 57 Table 27- Study 15982: Laboratory data ...... 59 Table 28- Study 15982: Race/Ethnicity and laboratory assessments ...... 61 Table 29- Study 15982: Age and laboratory assessments ...... 62 Table 30-Study 15982: Schedule of assessments (copied from submission) ...... 64 Table 31- Study 15982: AEs by PT (sorted by frequency, then alphabetical order) ...... 65

6 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table of Figures

Figure 1- 15982 Study Design ...... 19 Figure 2- Study 15982: Kaplan Meier curve, overall survival (ITT) ...... 34 Figure 3- Study 15982: FACT-Hep- means with 95% CI, Full Analysis Set (copied from submission) ...... 36 Figure 4- Study 15982: EQ-5D index score - means with 95% CI, Full Analysis Set (copied from submission) ...... 37

7 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

1 Recommendations/Risk Benefit Assessment

1.1 Recommendation on Regulatory Action

Approval is recommended, pending agreement on final labeling, for the use of regorafenib for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with Cb><4

-The applicant provided data establishing the safety and effectiveness of the product for the proposed indication as described under 21 CFR 314.70.

1.2 Risk Benefit Assessment

Benefit-risk summary and assessment Regorafenib is a small-molecule tyrosine kinase inhibitor that targets multiple kinases including VEGFR-2 and VEGFR-3. Regorafenib currently has indications in CRC and GIST and is being proposed for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with r Cb><4>j

HCC is associated with high morbidity and mortality. The 5-year OS for patients with advanced HCC in the U.S. is only approximately 12%. The only approved systemic therapy for the treatment of advanced HCC is sorafenib, a multikinase inhibitor with multiple targets including VEGF pathways. Based on prior studies, sorafenib improves median survival by approximately 2-3 months when compared to placebo. After HCC has progressed on sorafenib, the median OS is approximately 8 months and there are no approved systemic therapies for HCC after progression on sorafenib.

The efficacy of regorafenib for the treatment of patients with advanced HCC was demonstrated in Study 15982, a multicenter, randomized, placebo-controlled trial comparing once daily regorafenib to placebo in patients with HCC that had progressed on sorafenib. A total of 573 patients were randomized, 379 in the regorafenib arm and 194 in the placebo arm. Median overall survival was 10.6 months in the regorafenib arm and 7 .8 months in the placebo arm with a hazard ratio of 0.63 (95% Cl: 0.50, 0. 79) and an unstratified log-rank p-value of 0.0002.

The safety profile from Study 15982 was consistent with what is known about regorafenib for the treatment of metastatic CRC and GIST and was consistent with regorafenib being an oral VEGF-targeting tyrosine kinase inhibitor. Although adverse events were frequent in Study 15982, only about 10% of patients permanently discontinued regorafenib with the primary reason being an adverse event.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC The proposed indication for regorafenib following progression of Heer •><~

(b><">j indication for regorafenib, which is supported by Study 15982, is for the treatment of patients with (b)(4~ I

In summary, the approval is recommended based on a prolongation of overall survival with an acceptable toxicity profile, with which the oncology community has experience in managing. Analysis of condition Summary Regorafenib is an oral tyrosine kinase inhibitor that is proposed for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b)(4~

In the U.S. in 2016, an estimated 39,230 new diagnoses and 27, 170 deaths due to liver and intrahepatic bile duct cancers will have occurred, with the majority of these being HCC.1 In the U.S., the 5-year survival rate for HCC is approximately 12%.2

Advanced HCC is generally considered to be incurable and the goal of therapy is to prolong survival and improve quality of life . Systemic treatment is generally administered until disease progression or the toxicity of the therapy is deemed to be intolerable or detrimental to the quality of life. The only systemic therapy approved for the treatment of advanced HCC is the oral multikinase inhibitor sorafenib. Conclusion Advanced HCC is a progressive disease with a fatal outcome. The oral multikinase inhibitor sorafenib is the only approved systemic therapy for advanced HCC. Current Treatment Options Summary of evidence In 2007, sorafenib received regular FDA approval for the treatment of patients with advanced HCC. The median overall survival for patients with HCC that has progressed 3 on sorafenib, is approximately 7-9 months. -6 There are no systemic therapies approved to treat advanced HCC that has proqressed on sorafenib. Concl usion There are no approved systemic therapy options to treat advanced HCC that has proqressed on sorafenib. Therefore, reqorafenib addresses an unmet medical need. Clinical benefit Summary of evidence The efficacy of regorafenib for the treatment of patients with advanced HCC that has progressed on sorafenib was demonstrated in one multicenter, randomized, placebo- controlled trial, Study 15982.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

In Study 15982, patients with advanced HCC that had progressed on sorafenib were randomized 2:1 to receive regorafenib or placebo. The efficacy analysis of regorafenib was based on the ITT population of 573 patients with 379 patients in the regorafenib arm and 194 patients in the placebo arm. Patients were administered 160 mg of regorafenib or a matching placebo once daily.

Patient demographics were balanced between the two arms. The median age at randomization was 63 years and 66% of the patients had an ECOG performance status of O while 34% had an ECOG performance status of 1. Eighty eight percent of the patients were men. Forty one percent of the patients were Asian. Ninety eight percent of the patients had Child-Pugh class A cirrhosis and 2% had Child-Pugh class B. All patients had HCC that had progressed on sorafenib.

At the time of data cutoff, 233 (61 % ) of the patients in the regorafenib arm had died while 140 (72%) of the patients in the placebo arm had died. Sixty five patients (17.2%) in the regorafenib arm and 10 patients (5.2%) in the placebo arm were ongoing with study drug at the time of data cut off.

The primary endpoint of OS from randomization was met as median OS was 10.6 months in the regorafenib arm and 7.8 months in the placebo arm with a HR of 0.63 [95% Cl: 0.50, 0.79; p-value (unstratified log rank test) = 0.0002]. Subgroup analyses for OS were, in general, consistent with the OS in the ITT population.

The secondary endpoint of PFS was met based on both mRECIST for HCC and standard RECIST criteria. Using mRECIST, median time to disease progression was 3.1 months in the regorafenib arm compared to 1.5 months in the placebo arm [HR of 0.46 (95% Cl: 0.37, 0.56; p < 0.0001]. Using standard RECIST criteria, median time to disease progression was 3.4 months in the regorafenib arm compared to 1.5 months in the placebo arm [HR of 0.43 (95% Cl: 0.35, 0.52; p < 0.0001].

The secondary endpoint of ORR was met based on both mRECIST for HCC and standard RE Cl ST criteria. Using mRECIST criteria, the ORR was 10.6% in the regorafenib arm compared to 4.1 % in the placebo arm (p = 0.005) while using standard RECIST criteria, the ORR was 6.6% in the regorafenib arm compared to 2.6% in the placebo arm (p = 0.02). Conclusion Study 15982 was an adequate and well-controlled study that demonstrated that regorafenib for the treatment of advanced HCC in the 2nd_line setting resulted in a modest survival benefit. The use of regorafenib resulted in a median prolongation of survival of 2.8 months with median OS of 10.6 months and 7.8 months in the reqorafenib and placebo arms, respectively. Risk Summary of evidence The safety analysis was based on the safety population of Study 15982 with 37 4

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

patients in the regorafenib arm and 193 patients in the placebo arm. Overall, the incidence rates of adverse events of any grade were 100% in the regorafenib arm and 93% in the placebo arm which were comparable. The incidence rates of Grade 3-4 adverse events were 79% in the regorafenib arm compared to 56% in the placebo arm while the incidence rates of SAEs were 44% in the regorafenib arm compared to 47% in the placebo arm.

The most frequently reported (> 5%) Grades 3-4 adverse events in either arm were hypertension (15% and 5% in the regorafenib and placebo arms, respectively), palmar plantar erythrodysesthesia syndrome ( 12% and 1% in the regorafenib and placebo arms, respectively),increased AST (11% in each arm), hypophosphatemia (8% and 2% in the regorafenib and placebo arms, respectively), increased blood bilirubin (7% and 9%, in the regorafenib and placebo arms, respectively), increased lipase (7% and 2% in the regorafenib and placebo arms, respectively), and fatigue (6% and 4% in the regorafenib and placebo arms, respectively).

Grade 3-4 adverse events associated with VEGF inhibition and/or TKls were increased in the regorafenib arm and included hypertension (15% and 5% in the regorafenib and placebo arms, respectively), hemorrhagic shock (1 % and 0% in the regorafenib and placebo arms, respectively), and palmar-plantar erythrodysesthesia syndrome (12% and 1% in the regorafenib and placebo arms, respectively). The incidence of Grade 3-5 hypertension among regorafenib-treated patients was higher in patients <::: 65 years of age compared to younger patients (21 % in regorafenib-treated patients<::: 65 years compared to 9% in regorafenib-treated patients< 65 years).

Eighty eight deaths occurred during the treatment emergent adverse event period, 50 (13%) in the regorafenib arm and 38 (20%) in the placebo arm. The two arms were generally balanced with respect to the causes and, based on a focused review of narratives and brief case summaries, many of these deaths appeared to occur in the setting of clinical progression of disease.

The proportion of patients who discontinued study drug primarily due to an adverse event was similar between arms (10% and 9% in the regorafenib and placebo arms, respectively). Conclusion In summary, there were no new safety signals in Study 15982 and the safety profile was within what was expected for a regorafenib and for tyrosine kinase VEGF inhibitors, in general. Regorafenib appears to have a reasonable safety profile in this population and adverse events were generally manageable with supportive care or dose modification. Risk management The risks of regorafenib use in the treatment of advanced HCC are well known to prescribers and managed through product labeling. Additionally, this drug will be prescribed by oncologists who have specific training in the administration of anti neoplastic drugs and in the management of toxicities related to these drugs.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

1.3 Recommendations for Postmarket Risk Evaluation and Mitigation Strategies

No REMS were identified as necessary to approve this efficacy supplement.

1.4 Recommendations for Postmarket Requirements and Commitments

No new post-marketing commitments or requirements are recommended.

2 Introduction and Regulatory Background

Stivarga® (regorafenib) was approved in the U.S. on September 27, 2012, for the treatment of patients with metastatic CRC who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy, and, if KRAS wild type, an anti-EGFR therapy. On May 29, 2013, the U.S. package insert was expanded to include patients with locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have previously been treated with imatinib mesylate and sunitinib malate.

In this submission, the Applicant seeks approval for the following indication “the treatment of patients with hepatocellular carcinoma (HCC) who have been previously (b) (4) treated with

This application was submitted on October 30, 2016 (final sNDA rolling submission date) and the PDUFA goal date (priority review) is April 28, 2017. This review describes the efficacy and safety data supporting approval and the recommendation of the clinical reviewer.

2.1 Product Information

Stivarga® (regorafenib) is an oral bi-aryl urea compound that targets different receptor tyrosine kinases including vascular endothelial growth factor receptor 2 (VEGFR-2), vascular endothelial growth factor receptor (VEGFR-3), platelet-derived growth factor receptor (PDGFR), fibroblast growth factor receptor (FGFR), KIT proto-oncogene receptor tyrosine kinase (c-Kit), Raf-1 proto-oncogene, serine/threonine kinase (RAF1), and both wild-type and V600E-mutated B-Raf proto-oncogene, serine/threonine kinase (BRAF).7 These various kinases function in cellular metabolism, growth, and proliferation and are often inappropriately activated in tumor cells. Because of this, they are common targets of anti-cancer agents.

Stivarga is supplied as a 40mg tablet.

12 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

2.2 Currently Available Treatments for Proposed Indications

In its earlier, localized stages, HCC is often treated with locoregional therapies such as surgery (resection or orthotopic liver transplant), radiofrequency ablation (RFA), or transarterial chemoembolization (TACE). For patients with advanced or metastatic HCC and Child-Pugh A cirrhosis, first-line systemic treatment consists of sorafenib, an oral multikinase inhibitor, at 400 mg twice daily, based on results of the SHARP study8. In that study, 602 patients with advanced HCC were randomized to receive sorafenib 400 mg twice daily or placebo. The primary outcome was overall survival (OS) and this endpoint was met with a median OS of 10.7 months in the sorafenib arm versus 7.9 months in the placebo arm (hazard ratio of 0.69, 95% confidence interval 0.55 to 0.87, p < 0.001). In addition to the SHARP study, the effects of sorafenib for the treatment of patients with advanced HCC were assessed in 271 patients in a clinical trial conducted in China, South Korea, and Taiwan. In this trial, the median OS was 6.5 months compared to 4.2 months in the sorafenib and placebo arms, respectively (HR of 0.68, 95% confidence interval 0.50,0.93, p=0.014).9

There are currently no approved agents for the treatment of advanced or metastatic HCC that has progressed on sorafenib.

2.3 Availability of Proposed Active Ingredient in the United States

Stivarga® (regorafenib) was approved in the U.S. on September 27, 2012, for the treatment of patients with metastatic CRC who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy, and, if KRAS wild type, an anti-EGFR therapy. On May 29, 2013, the U.S. package insert was expanded to include patients with locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate.

2.4 Important Safety Issues With Consideration to Related Drugs

Though the exact adverse event profile varies among the different agents and also depends on patient- and disease-specific factors, the oral tyrosine kinase inhibitors (TKIs) generally have a characteristic pattern of toxicity that includes hypertension, proteinuria, skin reactions, thromboemboli, and diarrhea. Some of these effects are considered to be “on-target” effects of inhibiting particular pathways such as the VEGF (hypertension, proteinuria, thromboemboli) or EGFR (skin reactions) pathways.10 Similarly, drugs that predominantly target the VEGF pathway, regardless of whether they are small molecule TKIs or monoclonal antibodies, generally have a characteristic adverse event profile that includes hypertension, proteinuria, hemorrhage, thrombosis, fistula formation, and gastrointestinal perforation.11

13 Reference ID: 4081175 Cl inical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2 nd line HCC

2.5 Summary of Presubmission Regulatory Activity Related to Submission

On December 12, 2012, the Applicant submitted a new protocol for Study 15982: "Randomized, double-blind, placebo-controlled, multicenter phase Il l study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib" under IND 75642. On March 31 , 2015, a request for Orphan Drug Designation (ODD) was submitted for the treatment of patients with HCC and was granted on June 4, 2015.

On December 23, 2015, the Applicant requested a Type C meeting to discuss plans to submit a supplemental new drug application (sNDA) for Study 15982. On March 4, 2016, FDA issued Type C meeting written responses. The main issues addressed included FDA's recommendation that worse case imQutation be used for OS analysis in 4 the event of missin data. FDA also stated Cb>< -- A.aaitionally, FDA statea tnafffie aeterm1nat1on or Pnority Review woula e made at the time of sNDA filing.

On March 23, 2106, FDA issued an Advice/Information Request in response to the Applicant's March 18, 2016, letter requesting additional clarification regarding the statistical analysis plan. In FDA's March 23, 2016, letter, FDA recommended that Bayer use day 1 for the treatment arm and day 30 for the control arm as the worst case imputation approach for the OS analysis when data are missing. FDA also stated that the Applicant's plan to order the secondary endpoints for hierarchical testing using PFS followed by TTP was acceptable.

On July 21 , 2016, a Type B pre-sNDA meeting between the Applicant and FDA was held to discuss the efficacy and safety results of Study 15982 and to discuss the proposed indication for regorafenib. The Applicant proposed the following indication: "Stivarga is indicated for the treatment of Qatients with hepatocellular carcinoma (HCC) 4 who have been previously treated with Cb>< " . FDA stated that if the Fast Track designation is ranted it would need to incorQorate prior sorafenib treatment in the indication statement CbT<' FDA 4 also stated that the AQpl ican coula propose the Cb>< r 'prev1ously reated 4 with {bT( ") with justification at the time of original sNDA submission. At this meeting, it was also agreed that analyses regarding treatment post-progression would be considered exploratory. Discussions regarding the format of the summaries, safety updates, and the contents of the supplement were held , and the present submission follows the agreements reached during the meeting.

Fast Track designation was granted for the HCC development plan on July 28, 2016.

2.6 Other Relevant Background Information

HCC

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

An estimated 39,230 new diagnoses and 27,170 deaths due to liver and intrahepatic bile duct cancers occurred in the U.S. in 2016 with a male to female ratio of new diagnoses of 3:1 and of deaths of 2:1.1 The majority of these new cases and deaths are due to HCC, the most common primary cancer of the liver worldwide.12 Overall for HCC in the U.S., the 5-year survival rate is approximately 12%.2 Even when diagnosed and resected in its earlier stages, HCC carries a poor prognosis with a 5-year survival rate of approximately 50%.12 Additionally, the incidence of liver cancer has been increasing in both males and females in the U.S1 and, worldwide, liver cancer is one of the most common causes of cancer deaths and led to approximately 700,000 deaths in 2008 alone.13 Eastern and Southeastern Asia, Middle and Western Africa, Melanesia, and Micronesia/Polynesia have the highest incidence of liver cancer.13

The main risk factor for developing HCC is liver cirrhosis, found in 80-90% of patients with HCC.2,14 Causes of cirrhosis include hepatitis B virus (HBV) infection, hepatitis C virus (HCV) infection, alcohol use, non-alcoholic steatohepatitis, and, rarely, inherited metabolism disorders.14 In Asia and Africa, chronic HBV infection is the leading cause of HCC while in North America, Europe, and Japan, HCV infection is the leading cause.2,15

The Barcelona Clinic Liver Cancer (BCLC) staging system is widely used for HCC and is outlined below.16,17

BCLC Stage Tumor Characteristics Child-Pugh Score Single tumor or 3 tumors all A (Early) A-B < 3 cm B (Intermediate) Multinodular A-B Vascular invasion or C (Advanced) A-B extrahepatic spread D (End-Stage) Any C

Treatment Options in HCC Regarding treatment options for HCC, locoregional modalities include surgical resection and orthotopic liver transplantation. Surgical resection is often limited to patients without cirrhosis and who have early stage HCC. When surgical resection is attempted in patients with cirrhosis, good prognostic markers include small tumor (< 3 cm in diameter), normal total bilirubin level, and no portal hypertension (PH).2 Nevertheless, the overall 5-year risk of recurrence of HCC after resection is up to 70%.2 Orthotopic liver transplantation using the Milan criteria18 for patient selection typically has a 4-year overall survival (OS) rate of 85% and a 4-year recurrence free survival (RFS) rate of 92%2. Other local therapies, for disease confined to the liver, include radiofrequency ablation (RFA), transarterial chemoembolization (TACE), and radioembolization with yttrium-90.2,12

15 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

For more advanced HCC, the only approved standard systemic therapy is with the oral tyrosine kinase inhibitor (TKI) sorafenib, as discussed above in Section 2.2. There are currently no standard systemic therapy options available for patients with HCC that has progressed on sorafenib therapy. Based on the placebo arm of published reports of trials examining second-line agents for the treatment of HCC, the estimated median overall survival for patients with HCC that has progressed on sorafenib is approximately 7-9 months.3-6

3 Ethics and Good Clinical Practices

3.1 Submission Quality and Integrity

The submission was of adequate quality for the clinical review.

3.2 Compliance with Good Clinical Practices

The study reports contained in this sNDA included statements that the trials were conducted in accordance with the Declaration of Helsinki and The International Conference on Harmonization (ICH) GuidelineE6: Good Clinical Practice (GCP).

3.3 Financial Disclosures

Financial disclosure information was collected based on participation in Study 15982. Bayer stated that financial disclosure information was collected or due diligence was exercised to obtain the information and that each listed clinical investigator that was required to disclose if he/she had a proprietary interest in this product or significant equity in Bayer as defined in 21 CFR 54.2(b) did not disclose any such interests. Bayer also stated that no listed investigator was the recipient of significant payments of other sorts as defined in 21 CFR 54.2(f).

A total of 971 principal investigators (PIs) and sub-investigators (Sub-Is) signed financial disclosure forms. Of these 971 PIs and Sub-Is, 781 signed both initial and interim/end- of-study (EOS) forms and 190 signed only the initial forms because he or she left the trial site (118) or did not randomize any patients to the study (72). Six PIs and Sub-Is did not sign any forms and did not participate in the study. Additionally, Bayer provided Form 3455 for 4 PIs and Sub-Is who had disclosable information:

(b) (6) Two investigators at site , received approximately $91,000 for research grants, speaking fees, educational programs, or consulting from Bayer between 2013-2016. (b) (b) This site consented(6) patients,(6) of which received treatment.

16 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Reviewer’s Comment: The number of patients treated at this site constituted approximately (b) (6) % of the total patients treated. Additionally, Bayer examined the data from that site and did not find evidence of unusual outliers.

(b) (6) One investigator at site received approximately $36,000 in speaking and (b) (6) consulting fees and $50,000 in institutional grants from Bayer between 2014-2016. (b) patients were consented at this site and(6) received treatment.

(b) (6) One investigator at site received approximately $185,200 from Bayer for (b) (6) (b) consulting between 2013-2016. patients were consented and(6) received treatment.

Reviewer’s Comment: Bayer states that, although the details of the above payments to the investigator were not initially disclosed, the information from patients enrolled at this site was reviewed at the conclusion of the trial and no outliers in terms of safety or efficacy were identified. The number of patients treated at this site constituted approximately (b) (6) % of the total patients treated.

The Clinical Investigator Financial Disclosure Review Form is attached to this review.

4 Significant Efficacy/Safety Issues Related to Other Review Disciplines

4.1 Chemistry Manufacturing and Controls

Not applicable for this supplement.

4.2 Clinical Microbiology

Not applicable for this supplement.

4.3 Preclinical Pharmacology/Toxicology

Not applicable for this supplement.

4.4 Clinical Pharmacology

In addition to the clinical pharmacology information provided in previous regorafenib applications, Bayer provided data from three clinical pharmacology studies in this

17 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

application: Study 16653 (a pharmacokinetic study in cancer patients with renal impairment), Study 16674 (a drug-drug interaction study in cancer patients), and Study 16675 (a drug-drug interaction study in healthy volunteers). New clinical pharmacology information is also provided from Study 15982 (the RESORCE study). Sections 4.4.1, 4.4.2, and 4.4.3 briefly summarize clinical pharmacology information. Refer to the Office of Clinical Pharmacology review for additional analyses and conclusions regarding the data.

4.4.1 Mechanism of Action

In the application, Bayer stated that in in-vitro and in-vivo pre-clinical studies, regorafenib inhibits the growth, migration, and invasion of HCC cells and also induces apoptosis in HCC cells.

4.4.2 Pharmacokinetics

In the application, Bayer states that regorafenib is metabolized by multiple mechanisms, including glucuronidation and oxidation, to form two pharmacologically active metabolites, M-2 and M-5. Thus, this application includes data on unchanged regorafenib, M-2, and M-5.

Based on studies of patients with different races and ethnicities, Bayer states that dose adjustments based on race and ethnicity are not needed. Regarding liver function, Bayer states that the regorafenib exposure is similar between patients with normal liver function and patients with Child-Pugh class A liver impairment. There are limited data available regarding patients with Child-Pugh class B liver impairment and pharmacokinetic studies have not been performed on patients with Child-Pugh class C liver impairment. Based on renal impairment studies, Bayer states that patients with mild or moderate renal impairment demonstrated similar exposure of regorafenib compared to patients with normal renal function.

5 Sources of Clinical Data

5.1 Tables of Studies/Clinical Trials

One adequate and well-controlled study was used to support the expansion of the label in second line HCC, Study 15982 (PH-38451).

5.2 Review Strategy

The safety and efficacy analyses were based on the evaluation of Study 15982, “A randomized, double blind, placebo-controlled, multicenter phase III study

18 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib."

The first subject's first visit (FSFV) was on May 14, 2013, and the last subject's last visit (LSL V) on or before the data cutoff was on February 29, 2016. End of Study follow up is ongoing. The study enrolled patients from 152 clinical investigator sites in 21 countries in the U.S. , South America, Europe, Asia, and Australia. A total of 843 patients were enrolled. Twenty four centers in China, Spain, France, the United Kingdom, Hungary, Italy, Japan, South Korea, Russia, Taiwan, and the United States enrolled ten or more patients each (328, 39% ). Due to screen failures, a total of 573 patients were randomized. Of these, nine centers in China, France, Hungary, Italy, Japan, Russia, and Taiwan randomized ten or more patients each (127 patients, or 22%). Efficacy and safety analyses were conducted using the clinical database with a data cutoff date of February 29, 2016.

5.3 Discussion of Study 15982

Study 15982, "A randomized, double blind, placebo-controlled, multicenter phase Ill study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib"

The following protocol synopsis is based on the latest version of the protocol. Amendment #5 was approved December 1, 2015. Table 1 at the end of the protocol review summarizes major changes to the protocol since it was initiated.

Study Design Study 15982 was an international, randomized, double blind, placebo-controlled study. The study was managed by Bayer HealthCare AG.

The study compared the efficacy and safety of second-line regorafenib versus placebo in patients with HCC that had progressed on sorafenib. The following figure shows the study design. Figure 1- 15982 Study Design

R A Regorafenib 160 mg daily, maximum 3 weeks on and 1 week off of 10 weeks N after last D dose of 0 Progressive sorafenib M disease while I Disease progression/death on sorafenib z A T I Placebo 160 mg daily, 3 0 weeks on and 1 week off N 19 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Patients were randomized in a 2:1 ratio to receive either regorafenib or placebo using a computer-generated randomization list. The randomization number for each eligible patient was then provided to the investigator through an interactive voice response system (IVRS). Patients were stratified based on geographical region (Asia versus rest of world), ECOG performance status (0 versus 1), AFP level (< 400 ng/mL versus ≥ 400 ng/mL) extrahepatic disease (presence vs absence), and macrovascular invasion (presence versus absence). Study treatment was administered until progressive disease (PD) based on RECIST v1.1 or based on mRECIST for HCC patients, clinical progression (worsening ECOG performance status to ≥ 3 or symptomatic deterioration including increasing LFTs), death, unacceptable toxicity, or decision by patient or treating physician. Of note, the protocol allowed for continuation of treatment beyond PD if the treating physician felt the treatment was providing clinical benefit.

Objectives The primary objective of this study was to determine overall survival (OS).

Secondary objectives were analyses of time to progression (TTP), progression free survival (PFS), objective tumor response rate (ORR), and disease control rate (DCR) defined as complete response (CR) + partial response (PR) + stable disease (SD).

Tertiary objectives were analyses of duration of response (DoR), duration of stable disease, health related quality of life and utility values, pharmacokinetics (PK), and biomarker evaluation.

Safety objectives were the analysis of adverse events, physical examination and vital signs, laboratory assessments, 12-lead electrocardiograms (ECG), Child-Pugh status, and ECOG performance status.

Study Population (modified for brevity)

Inclusion Criteria: 1. Patients with histological confirmation of HCC or with non-invasive diagnosis of HCC based on the American Association for the Study of Liver Diseases (AASLD) criteria in patients with confirmed cirrhosis. 2. Patients must have HCC that has progressed on prior treatment with sorafenib. Progression on sorafenib was defined as radiologic progression based on the radiology charter. Patients must have tolerated treatment with sorafenib (defined as not less than 20 days on at least 400 mg daily within the last 28 days prior to stopping the sorafenib). Patients had to be randomized within 10 weeks after the last dose of sorafenib.

20 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

3. BCLC stage category B or C that cannot benefit from treatments with established efficacy. 4. Liver function status Child-Pugh class A as calculated during the screening period. 5. Any local or locoregional therapy of intrahepatic tumors must have been completed ≥ 4 weeks before 1st dose of study medication. 6. Adequate organ function including: a. Total bilirubin ≤ 2 mg/dL b. ALT and AST ≤ 5 X upper limit of normal (ULN) c. Serum creatinine ≤ 1.5 X ULN d. Glomerular filtration rate (GFR) ≥30 mL/min/1.73 m2 e. Platelet count ≥ 60,000/mm3 7. ECOG performance status 0-1. 8. Age ≥ 18 years. 9. Signed written informed consent. 10. Agreement for men and women to use adequate contraception methods.

Exclusion Criteria: 1. Prior systemic treatment for HCC except sorafenib. 2. Sorafenib treatment within 2 weeks of randomization. 3. Permanent discontinuation of sorafenib therapy due to sorafenib-related toxicity. 4. Candidate for liver transplantation. 5. Patients with large esophageal varices at risk for bleeding and that were not being treated with standard medical treatment. 6. Ascites not controlled with diuretic or paracenteses. 7. Ascites or pleural effusion causing respiratory compromise. 8. Clinically significant bleeding event ≥ Grade 3 within 30 days of randomization. 9. Uncontrolled hypertension defined as systolic blood pressure >150 mmHg or diastolic pressure >90 mmHg despite optimal medical management. 10. Arterial or venous thrombotic or embolic events within 6 months before 1st dose of study medication. 11. Persistent proteinuria ≥ Grade 3. 12. Non-healing wound, ulcer, or bone fracture. 13. Hepatitis B infection unless no active viral replication is present. 14. Hepatitis C infection that requires antiviral treatment. 15. NYHA CHF Class ≥ 2. 16. Unstable angina or myocardial infarction within 6 months of randomization. 17. Cardiac arrhythmias requiring anti-arrhythmic medication, excluding beta- blockers and digoxin. 18. Interstitial lung disease with ongoing symptoms at screening. 19. Known history of or presence of a symptomatic CNS tumor. 20. Known history of HIV infection.

21 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

21. Known hypersensitivity to any of the study drugs, study drug classes, or excipients in the formulation. 22. Inability to swallow or any malabsorption condition. 23. Pregnancy or breast feeding.

Reviewer’s Comment: The protocol adequately selected for a patient population with mild cirrhosis (98% of the patients had Child-Pugh A class cirrhosis) but did not enroll patients with Child-Pugh class B or C cirrhosis (except for the 2% of patients with class B cirrhosis). Thus, the safety and efficacy in these groups of patients is unknown. The primary objective of overall survival is appropriate for this population of patients.

Treatment Plan

Patients randomized to receive regorafenib received 160 mg (4 x 40-mg tablets) orally daily for 3 weeks of every 4 week cycle and also received best supportive care (BSC). Patients randomized to receive placebo received 160 mg (4 x 40-mg) of matching placebo tablets on the same schedule as the patients receiving regorafenib and also received BSC. Study medication was delayed or reduced based on adverse events graded using NCI CTCAE v4.03, according to pre-defined guidelines. Dose levels were as follows:

Does Level 0: Standard dose, 160 mg orally daily Dose Level -1: 120 mg orally daily Dose Level -2: 80 mg orally daily

No more than 2 dose reductions were allowed. Intra-patient dose re-escalation was allowed.

Dose modifications and delays for Grade 3-4 AST and/or ALT increases related to study drug were as follows:

Grade 3, 1st occurrence: Interrupt treatment until ≤ Grade 2 or baseline and then resume at one dose level lower.

Grade 3, 2nd occurrence: Interrupt treatment until ≤ Grade 2 or baseline and then resume at a second dose level lower.

Grade 3, 3rd occurrence: Discontinue study drug.

Grade 3 with ALT or AST > 8 x ULN and a concomitant increase in total bilirubin level to any degree, 1st occurrence: Interrupt treatment until ≤ Grade 2 or baseline and then resume at one dose level lower or consider discontinuing study drug.

22 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Grade 3 with ALT or AST > 8 x ULN and a concomitant increase in total bilirubin level to any degree, 2nd occurrence: Discontinue study drug.

Grade 4: Discontinue study drug.

Dose modifications and delays for Grade 2-3 palmar-plantar erythrodysesthesia syndrome (PPES)/hand-foot skin reaction (HFSR) related to study drug were as follows:

Grade 2, 1st occurrence: Institute supportive measures and consider decreasing dose by one level. If no improvement, interrupt study drug for at least seven days until resolves or improves to Grade 1.

Grade 2, no improvement after seven days interruption or 2nd occurrence: Interrupt study drug until resolves or improves to Grade 1 and then resume study drug at one dose level reduced.

Grade 2, 3rd occurrence: Interrupt study drug until resolves to Grade 1 and the resume study drug at 2 dose levels reduced total.

Grade 2, 4th occurrence: Discontinue study drug.

Grade 3, 1st occurrence: Institute supportive measures and interrupt study drug for at least seven days until resolves or improves to Grade 1 and then resume study drug at one dose level reduced.

Grade 3, 2nd occurrence: Institute supportive measures and interrupt study drug for at least seven days until resolves or improves to Grade 1 and then resume study drug at two dose levels reduced total.

Grade 3, 3rd occurrence: Discontinue study drug.

Dose modifications and delays for treatment-emergent hypertension were as follows:

Grade 1: Consider increasing blood pressure monitoring.

Grade 2: If not symptomatic, continue study drug but treat with anti-hypertensives for goal diastolic blood pressure (BP) ≤ 90 mmHg. If symptomatic, interrupt study drug until symptoms resolve and diastolic BP ≤ 90 mmHg.

Grade 3: Interrupt study drug until any symptoms resolve and goal diastolic BP ≤ 90 mmHg and then resume at the same dose level. If BP is not controlled with anti-hypertensive, reduce dose by one dose level. If Grade 3 recurs despite dose reduction and anti-hypertensive therapy, reduce study drug by another dose level.

23 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Grade 4: Discontinue study drug.

Dose modifications and delays for all other Grade 3-4 adverse events related to study drug were as follows:

Grade 3: Interrupt study drug until ≤ Grade 2 and reduce dose by one dose level. Grade 4: Interrupt study drug until ≤ Grade 2 and consider discontinuation of study drug.

Efficacy assessments

Tumor measurements, response, and disease progression were assessed using RECIST v1.1 and mRECIST for HCC criteria every 6 weeks +/- 7 days after baseline. All post-baseline imaging assessments were required to be performed using the same technique (including modality and slice thickness) as the baseline scan.

The study schedule of assessments is in the Appendices section (Table 30).

Safety

Adverse events were evaluated using CTCAE v4.03 and adverse events of all grades were collected, regardless of relationship to study drug or placebo. Hepatic failure, hepatobiliary disorders (Grades 4 and 5), and bleeding/hemorrhagic events (Grade ≥3) were considered adverse events of special interest (AESI). A treatment-emergent AE (TEAE) was defined as any event arising or worsening after initiation of study drug until 30 days after the last study drug intake. “Common” TEAEs were defined as AEs with at least a 5% total incidence rate of any grade.

The study schedule of assessments is in the Appendices section of this review (Table 30).

Withdrawal criteria:

• Patient request

• Investigator assessment that continuation would be harmful to patient

• Substantial non-compliance by patient

• Illicit drug use by the patient that could increase toxicity or confound assessments

24 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

• Severe allergic reaction or any other potential adverse reaction felt to warrant discontinuation

• Development of intercurrent illness or situation which would impact assessments and study endpoints

• Progressive disease (Of note, if an investigator felt that a patient could receive clinical benefit from continued treatment, the patient was allowed to remain on study.)

• Clinical progression

• Development of a second malignancy

• Loss of patient to follow up

• Interruption of study drug administration for more than 28 consecutive days

• Need for more than two dose reductions

• Pregnancy

Of note, if an investigator felt that a patient could receive clinical benefit from continued treatment, the patient was allowed to remain on study.

Reviewer’s Comment: The efficacy and safety monitoring plans were adequate for this population.

Statistical Issues

Primary endpoint The primary endpoint was overall survival (OS), performed on the intention to treat (ITT) population. OS was defined as the time from randomization to death due to any cause. Patients were stratified based on geographical region (Asia versus rest of world), ECOG performance status (0 versus 1), AFP level (< 400 ng/mL versus ≥ 400 ng/mL) extrahepatic disease (presence vs absence), and macrovascular invasion (presence versus absence).

Secondary endpoints Secondary endpoints were:

Time to progression (TTP) was defined as the time in days from randomization to radiological or clinical progression. In this study, clinical progression was defined

25 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

as worsening of the ECOG performance status to ≥ 3 or symptomatic deterioration including an increase in liver function tests (LFTs).

Progression free survival (PFS) was defined as the time in days from randomization to progression (radiological or clinical) or death due to any cause.

Objective tumor response rate (ORR) was defined as the rate of patients with complete response (CR) or partial response (PR) over all randomized patients. Patients who discontinued without an assessment were considered non- responders for the analysis.

Disease control rate (DCR) was defined as the rate of patients with CR, PR, or stable disease (SD) over all treated patients.

Tumor response and disease progression were evaluated based on RECIST v1.1 and also mRECIST for HCC.

Tertiary endpoints Tertiary endpoints were:

Duration of response was defined as the time from first documented objective response to disease progression or death and was evaluated using RECIST v1.1 and mRECIST for HCC criteria.

Duration of stable disease was defined as the time from randomization to the date that disease progression or death was first documented and was evaluated using RECIST v1.1 and mRECIST for HCC criteria.

Pharmacokinetics (PK) of regorafenib

Health related quality of life (HRQoL) using the FACT-Hep and the EQ-5D instruments.

Biomarker evaluation on patient blood samples and tumor tissue were planned to assess the mechanisms of action of regorafenib in patients with HCC.

For this study, the full analysis set (FAS) (i.e., the intent-to-treat analysis set) was defined as all randomized patients. The safety analysis set (SAF population) comprised all randomized patients who received at least one dose of study medication. Patients who were withdrawn from the study were not replaced.

Regarding interim analyses, one formal interim futility analysis was performed. A second formal interim analysis, for OS, was originally planned but this requirement was

26 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

removed in amendment 4 of the protocol and was not performed (see Protocol Amendment section, below).

The sample size was calculated based on the OS endpoint with a targeted improvement of a 43% increase in median OS compared to placebo and an associated hazard ratio (HR) of 0.7. Approximately 370 events, assuming a one-sided a of 0.025, were required to achieve a 43% improvement in median OS with a power of 90% and a randomization of 2:1 between regorafenib and placebo. Ultimately, 843 patients were enrolled and 573 patients were randomized.

Reviewer's Comment: The primary endpoint of overall survival is an appropriate endpoint to establish benefit. There are currently no approved therapies for the treatment of patients with advanced or metastatic HCC that has progressed on sorafenib therapy.

Protocol amendments

Study 15982 had 5 protocol amendments, all of which were in effect prior to the data cutoff for the report submitted to FDA for review. Table 1 summarizes the major changes in each protocol amendment.

Table 1- Study 15982: Protocol Amendments Amendment 1 (2 May 2013) v 2.0 • Inclusion criteria were modified to allow patients who had demonstrated progression in previously-treated tumor lesions. • The requirement for the assessment of esophageal varices by endoscopy within 6 months and 12 months of the start of the study was modified to require that endoscopy be peifo1med as per local standard of care. • Text was modified stipulating that patients should start treatment no later than 3 days after randomization. • The requirement for dose adjustments or intenuptions in response to increasing bilirubin levels was removed because it was felt that isolated elevations ofbilirubin levels without concomitantly elevated transaminases did not require strict dose modification. • Text was added stating that strong inhibitors of CYP3A4 activity should be avoided . Amendment 2 (13 December 2013) v 3.0 • The time of randomization of patients was to take place within 10 weeks of a patient's last treatment with sorafenib (changed from 8 weeks to accommodate patients transfeITed from outside sites). • In the section on radiological assessment text was added clarifying how tumor evaluation for HCC should be conducted using RECIST 1.1 and mRECIST.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

• A section entitled "Adverse events of special interest" was added. Amendment 3 (11 November 2014) v 4.0 • The number of patients to be recrnited into the study was increased from 530 to 560 to enable 150 patients from China to be recruited while maintaining the 40% cap for Asian patients. • A change was included so that patients who had had previous intrahepatic intraarterial chemotherapy with lipiodol could be emolled in the study. Amendment 4 (2 November 2015) v 5.0 • The requirement for the 2nd interim analysis was removed because slower than expected accrnal in China would have led to it being conducted prior to full patient accrual. Amendment 5 (1 December 2015), v 6.0 • Info1mation was added regarding interactions between regorafenib and neomycin, breast cancer resistant protein (BCRP), UGTIAl, UGT1A9, P-glycoprotein substrates, and bile salt sequestering agents.

6 Review of Efficacy

Efficacy Summary

Efficacy analyses were based on the ITT population which consisted of 573 patients (379 patients in the regorafenib arm and 194 patients in the placebo arm).

Patient demographics were balanced between the two arms. The median age at randomization was 64 years in the regorafenib arm and 62 years in the placebo arm. Eighty eight percent of the patients were men in both arms. This percentage of men reflects the higher proportion of male patients in the HCC population although estimates described in literature range between 2:1 and 4:1 male:female.14 Although data on race was missing for 20% and 23% of the patients in the regorafenib and placebo arms, respectively (due to laws prohibiting the collection of these data in particular countries), the two arms appeared balanced with respect to race and each arm had 38% of patients from Asia and 62% from the rest of the world (ROW).

At the time of data cutoff, 233 (61 %) of patients in the regorafenib arm had died and 140 (72%) of patients in the placebo arm had died. The study met the primary endpoint of demonstrating an improvement in overall survival in the regorafenib arm. The median OS was 10.6 months in the regorafenib arm compared to 7.8 months in the placebo arm with a hazard ratio (HR) of 0.63 [95% Cl: 0.50, 0.79; p-value (unstratified log-rank test) = 0.0002]. The median OS time was increased by 2.8 months. Subgroup analyses were exploratory but were, in general, consistent with the OS in the ITT population.

The secondary endpoints of PFS and ORR were also met when assessed by either modified RECIST for HCC criteria (mRECIST) or standard REC IST criteria. The

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

mRECIST for HCC criteria led to a more conservative estimation of PFS [HR = 0.46 (0.37, 0.56) using mRECIST and HR = 0.43 (0.35, 0.52) using RECIST] while standard RECIST criteria led to a more conservative estimate of ORR [point estimates of 10.6% and 4.1% in the regorafenib and placebo arms, respectively, using mRECIST and 6.6% and 2.6% for the regorafenib and placebo arms, respectively, using RECIST].

The mRECIST criteria for HCC were proposed in 2008 and aimed to improve the radiologic assessment of HCC, particularly in the setting of locoregional or molecularly- targeted therapies.19 These modified criteria use the arterial uptake in contrast- enhanced imaging techniques to assess for viable (versus necrotic) tumor tissue.19 Increasingly, clinical studies for patients with HCC are reporting results using mRECIST in addition to or instead of RECIST. Related to this concept, a modified version of mRECIST that specifically measures the change in arterial enhancement of HCC lesions also has been proposed.20

In Study 15982, the primary endpoint was OS and this endpoint was met. PFS and ORR were secondary endpoints although the Office of Biostatistics has recommended that ORR be exploratory because it was not included in the pre-specified hierarchical test order for secondary endpoints. For PFS, mRECIST criteria led to a more conservative estimation compared to the estimation by standard RECIST criteria though both were similar and both favored the regorafenib arm.

6.1 Indication

Currently, regorafenib is indicated for the treatment of patients with 1) metastatic colorectal cancer (CRC) who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy, and, if RAS wild-type, an anti-EGFR therapy or 2) locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate.

This supplement aims to expand the indication by adding “for the treatment of patients (b) (4) with hepatocellular carcinoma (HCC) who have been previously treated with ”.

6.1.1 Methods

The safety and efficacy analyses were centered on the evaluation of one trial, Trial 15982 (PH-38451), “A randomized, double blind, placebo-controlled, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib.”

29 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

6.1.2 Demographics

Study 15982 enrolled 843 patients at 152 sites located in 21 countries in the U.S., South America, Europe, Asia, and Australia, 270 of which were screen failures. Thus, 573 patients were ultimately randomized. The first patient's first visit was in May 2013 and the last patient's last visit on or before the data cutoff was in February 2016. Nine centers in Ch ina, France, Hungary, Italy, Japan, Russia, and Taiwan randomized ten or more patients each ( 127 patients, or 22% ). Efficacy and safety analyses were conducted using the clinical database with a data cutoff date of February 29, 2016.

Patient demographics were generally balanced between the two treatment arms (Table 2). Median age at randomization was 62 years in the placebo arm and 64 years in the regorafenib arm. Eighty eight percent of the patients were men in both arms which reflects the higher proportion of male patients in the HCC population although estimates described in literature range between 2:1 and 4:1 male:female. 14 Regarding race, not all of the participating countries require or allow reporting of race but of those that report, the two arms were generally balanced with 40.2% of the patients in the placebo arm being Asian and 41.2% of the patients in the regorafenib arm being Asian. A total of 38% of patients were enrolled from Asian study sites.

Table 2- Study 15982: Demographics (ITT population) no. of PL(% ofN) no. ofRegorafenib (% ofN) N= 194 N=379 Gender Male 171 (88) 333 (88) Female 23 (1 2) 46 (1 2) Alle (years) Range (years) 23-83 19-85 Mean(SD) 61.1 (1 1.6) 61.8 (1 2.4) Median 62 64 Yow1ger than 65 years 116 (60) 199 (53) Race Asian 78 (40) 156 (41) Black or African American 2 (1) 6 (2) White 68 (35) 138 (36) Multiple 1 (0.5) 2 (0.5) Not reported 45 (23) 77 (20) Reefon Asia 73 (38) 143 (38) Rest of the world 121 (62) 236 (62)

Initial disease characteristics were similar and balanced between treatment arms (Table 3). Sixty six percent of patients had ECOG performance status of O in both arms. However, regarding randomization based on ECOG performance status, data input for randomization differed for 19 patients (3.3% of the 573 patient ITT population). A

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

sensitivity analysis performed by the Office of Biostatistics did not reveal a significant impact due to this difference.

The etiology of HCC was similar in both arms and was due to hepatitis B or C or alcohol use in the majority of patients in both arms. A patient could have more than one etiology so the sum of the individual etiologies is not identical to the total number of patients in a given treatment arm.

For Barcelona Clinic Liver Cancer (BCLC) classification and Child-Pugh score at study entry, the placebo and regorafenib arms were similar with 89% and 86% of patients classified as BCLC stage C and 97% and 98% of patients classified as Child-Pugh A, in the placebo and regorafenib arms, respectively. The eligibility criteria for Study 15982 stipulated that only patients with Child-Pugh A classification (and BCLC stages B or C) at baseline are eligible. The enrollment of patients with Ch ild-Pugh score B or BCLC stage A are discussed in Section 6.1 .3 Subject Disposition.

Alpha-fetoprotein (AFP) values for the two arms are also recorded in Table 3. Values were not available for 3 (1.5%) and 5 (1.3%) patients in the placebo and regorafenib arms, respectively. However, regarding randomization based on AFP group (;::: 400 ng/ml versus < 400 ng/ml), data input for randomization differed for those 8 patients with missing AFP values (1.4% of the 573 ITT population). A sensitivity analysis performed by the Office of Biostatistics did not reveal a significant impact due to this difference.

Regarding macrovascular invasion and/or extrahepatic disease, the placebo arm had a smaller percentage of patients with macrovascular invasion only but a larger percentage of patients with extrahepatic disease only compared to the regorafenib arm. Both conditions were present in 20% and 19% of the placebo and regorafenib arms, respectively while the placebo arm had a smaller percentage (16%) of patients with neither condition, compared to the regorafenib arm (20%).

Based on the medical history dataset, 429 patients had a medical history of cirrhosis (75% of the 573 patients in the ITT population). Of these 429 patients, 7% and 6% of the patients had documented ascites and 18% and 13% of the patients had documented esophageal varices at the time of study entry, in the placebo and regorafenib arms, respectively.

Table 3-Study 15982: Disease characteristics (ITT population) no. of PL(% ofN) I no. ofRegorafenib (% ofN) I N=194 N=379 ECOGPS 0 I 129 (66) I 251 (66) 1 I 65 (34) I 128 (34) Etiolo2V of HCC

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2 nd line HCC

no. of PL (% of N) no. ofRegorafenib (% ofN) N=194 N=379 Hepatitis B 73 (38) 143 (38) Hepatitis C 41 (21) 78 (21) Alcohol use 55 (28) 90 (24) Other (not Hepatitis or Alcohol) 46 96 BCLC stage at stu dy entry A (early) 0 (O) 1 (0.3) B (intennediate) 22 (11) 53 (14) C (advanced) 172 (89) 325 (86) Child-Pugh score at study enh-y A5 118 (61) 244 (64) A6 70 (36) 129 (34) B7 5 (3) 5 (1) B8 1 (0.5) 0 (O) Missing data 0 (O) 1 (0.3) Alpha-fetoprotein (AFP) group :::: 400 ng/mL 85 (44) 158 (42) < 400 ng/mL 106 (55) 216 (57) Missing data 3 (1.5) 5 (1 .3) Alpha-fetoprotein (AFP) ng/mL Meanng/mL 12622 13508 Median ng/mL 234 183 Missing data (number of patients) 3 5 Macrovascular invasion Present 54 (28) 110 (29) Absent 140 (72) 269 (71) Extrahepatic disease Present 147 (76) 265 (70) Absent 47 (24) 114 (30) Macrovascular invasion and Extrahepatic disease Macrovascular invasion only 15 (8) 39 (10) Extrahepatic disease only 108 (56) 194(51) Both present 39 (20) 71 (19) Neit11er present 32 (16) 75 (20) Presences of ascites at study entry no. of PL (% of N) no. of Regorafenib (% of N) N=144 N =285 (patients with cirrhosis) (patients with cirrhosis) Yes 10 (7) 17 (6) No 134 (93) 268 (94) Presence of esophageal varices at no. of PL (% of N) no. of R egorafenib (% of N) study entry N =144 N =285 (patients with cirrhosis) (patients with cirrhosis) Yes 26 (18) 37 (13) No 118 (82) 248 (87)

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

6.1.3 Subject Disposition

Of the randomized patients, 194 were assigned to the placebo (PL) arm and 379 were assigned to the regorafenib arm with 573 patients in the full analysis set (FAS). However, 5 patients randomized to regorafenib did not receive treatment and 1 patient randomized to placebo did not receive treatment. The safety population included the 567 patients who received at least one dose of study medication: 374 patients who received regorafenib and 193 patients who received placebo.

Of the 567 patients who started study drug, 65 patients (17.2% of the 379 randomized patients) in the regorafenib arm and 10 patients (5.2% of the 194 randomized) in the placebo arm were ongoing with study drug at the time of data cut off.

6.1.3.1 Protocol violations and deviations

Bayer reported 1 protocol violation in the regorafenib arm (0.3%) and 3 protocol violations in the placebo arm (1.5%) that led to termination of treatment.

The only major protocol deviation was defined as randomization of a patient who then does not receive treatment. As discussed above, 5 patients in the regorafenib group (1.3%) and 1 patient in the placebo group (0.5%) had major protocol deviations.

Minor protocol deviations were reported in 100% of the patients in both arms of the trial and those that were considered important largely involved excluded concomitant medication treatment, inclusion/exclusion criteria not met but patient received treatment, and patient met withdrawal criteria but remained in the treatment phase. Table 8-2 in the Clinical Study Report (CSR) summarizes these data and the two treatment arms were similar with respect to the reasons for minor protocol deviations. Regarding patients not meeting inclusion/exclusion criteria but still entering the treatment phase, 5 patients in the regorafenib arm (1.3%) and 6 patients in the placebo arm (3%) had Child-Pugh class B cirrhosis despite the protocol excluding patients with Child-Pugh class B or C. Additionally, 1 patient in the regorafenib arm (0.3%) had Child-Pugh class information missing. Similarly, 1 patient in the regorafenib arm (0.3%) had BCLC stage A disease despite the protocol excluding patients with stage A disease.

Despite these violations and deviations, their number and nature of the violations would not appear to invalidate the analysis of overall survival.

6.1.4 Analysis of Primary Endpoint

The primary endpoint of Study 15982 was overall survival (OS). Overall survival was defined as the time from randomization to death due to any cause. At the time of data cutoff, 233 (61%) of patients in the regorafenib arm had died and 140 (72%) of patients

33 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

in the placebo arm had died. Table 4 summarizes the overall survival results in the ITT population. The median OS was 10.6 months in the regorafenib arm compared to 7.8 months in the placebo arm. The median time was increased by 2.8 months in the regorafenib arm.

Table 4- Study 15982: Overall survival results (ITT) PL; 0(%) Regonfenib; n (%) N=194 N=379 Alive 54 (28) 146 (39) Death 140 (72) 233 (61) Time to event (95% CI) 7.8 (6.3, 8.8) 10.6 (9.1, 12.1) HR(95% CI) 0.63 (0.50, 0.79) p-value (unstratified log rank test) 0.0002

The Kaplan-Meier plot (Figure 2) showed separation of the curves in favor of the regorafenib arm. The curves appeared to separate early and continued to be separate at two years.

Figure 2- Study 15982: Kaplan Meier curve, overall survival (ITT)

0 2 4 6 8 10U l•16 18202224 26 283032

Mooths

- Ftcgorei:fuiibl60mg

6.1.5 Analysis of Secondary Endpoints

Secondary endpoints in this study were analyses TTP, PFS, ORR, and OCR (b)(4 ORR and --~~~~~~~~~~~~~~~~~~~~~~~~~~~----- PFS, assessed by both mRECIST for HCC and RECIST, were analyzed in the Office of Biostatistics review and a summary of their analyses are summarized below (Table 5

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

and Table 6). The Office of Biostatistics recommends that the ORR endpoint be considered exploratory because ORR testing was not included in the pre-specified hierarchical test order for secondary endpoints.

Table 5- Study 15982: PFS results (ITT population) PFS by mRECIST PL; n(o/o) Regorafenib; n (%) N=194 N=379 # PFS events 181 (93) 293(77) # progressive disease 173 274 # deaths 8 19 mPFS (95% CI) in months 1.5 (1.4, 1.6) 3.1 (2.8, 4.2) HR(95%CI) 0.46 (0.37' 0.56) p-value < 0.0001 PFS by RECIST PL; n(o/o) Regorafenib; n (%) N= 194 N=379 # PFS events 184 (95) 288 (76) # progressive disease 175 270 #deaths 9 18 mPFS (95% Cl) in months 1.5 (I .4, 1.5) 3.4 (2.9, 4.2) HR(95% Cl) 0.43 (0.35, 0.52) p-value < 0.0001

Table 6- Study 15982: ORR results (ITT population) ORR by mRECIST PL; n(o/o) Regorafenib; n (%) N=194 N=379 Response Rate (95% en 4.1 % (1.8%, 8.0%) 10.6% (7.6%, 14.1%) Responders (CR+PR) 8 40 CR 0 2 PR 8 38 p-value 0.004728 Median duration of 2.7 (1.9, NE) 3.5 (1.9, 4.5) response (months) ORR by RECIST PL; n(o/o) Regorafenib; n (%) N= 194 N=379 Response Rate (95% en 2.6% (0.8%, 5.9%) 6.6% (4.3%, 9.6%) Responders (CR+PR) 8 25 CR 0 0 PR 8 25 p-value 0.019991 Median duration of 5.6 (2.3, NE) 5.9 (1.4, 8.4) response (months)

6.1.6 Other Endpoints

Duration of response analyses are included in Table 6.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

FACT-Hep and EQ-5D patient-reported outcomes (PRO) instruments were administered to patients during the trial. These instruments were self-administered at Cycle 1-Day 1, at every cycle, and then at the end-of-study visit. FACT-Hep scores range from 0 to 180 with higher scores correlating with higher quality of life. EQ-5D scores range from -0.59 to 1.0 with higher scores correlating with higher health states.

Figure 3- Study 15982: FACT-Hep- means with 95% CI, Full Analysis Set (copied from submission)

36 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Figure 4- Study 15982: EQ-5D index score - means with 95% CI, Full Analysis Set (copied from submission)

As demonstrated in Figure 3 and Figure 4, differences between arms using the FACT- Hep and EQ-5D did not meet minimally important difference thresholds and, thus, did not identify any clinically meaningful differences between the two arms.

6.1.7 Subpopulations

Patients were stratified based on geographical region (Asia versus rest of world), ECOG performance status (0 versus 1), AFP level (< 400 ng/mL versus ≥ 400 ng/mL) extrahepatic disease (presence vs absence), and macrovascular invasion (presence versus absence). The Office of Biostatistics analyzed OS based on these subgroups and most of the OS analyses for these subgroups were consistent with the primary analysis. All of these subgroup analyses are considered exploratory.

37 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

6.1.8 Analysis of Clinical Information Relevant to Dosing Recommendations

Not applicable.

6.1.9 Discussion of Persistence of Efficacy and/or Tolerance Effects

Not applicable.

6.1.10 Additional Efficacy Issues/Analyses

Not applicable.

7 Review of Safety

Safety Summary The safety analysis was conducted based on data from the safety population of Study 15982 (374 patients in the regorafenib arm and 193 patients in the placebo arm). Overall, incidence rates of adverse events of any grade (100% in the regorafenib arm and 93% in the placebo arm) were comparable between the treatment arms. The high incidence of adverse events in the placebo arm reflects the morbidity of advanced HCC. The incidence of Grade 3-4 adverse events was higher in the regorafenib arm compared to the placebo arm (79% compared to 56%).

The reported incidence of death related to adverse events was lower in the regorafenib arm (13% compared to 20%). The most frequently reported causes of Grade 5 events were general physical health deterioration (23 in the regorafenib arm and 16 in the placebo arm) and hepatic failure (3 in the regorafenib arm and 5 in the placebo arm).

The most frequently reported (> 5%) Grades 3-4 adverse events in either arm were hypertension (15% and 5% in the regorafenib and placebo arms, respectively), PPES (12% and 1% in the regorafenib and placebo arms, respectively), increased AST (11% in each arm), hypophosphatemia (8% and 2% in the regorafenib and placebo arms, respectively), increased blood bilirubin (7% and 9%, in the regorafenib and placebo arms, respectively), increased lipase (7% and 2% in the regorafenib and placebo arms, respectively), and fatigue (6% and 4% in the regorafenib and placebo arms, respectively).

Grade 3-4 adverse events associated with VEGF inhibition and/or TKIs were increased in the regorafenib arm and included hypertension (15% and 5% in the regorafenib and placebo arms, respectively), hemorrhagic shock (1% and 0% in the regorafenib and placebo arms, respectively), and PPES (12% and 1% in the regorafenib and placebo arms, respectively). The incidence of Grade 3-5 hypertension among regorafenib-

38 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

treated patients was higher in patients ≥ 65 years of age compared to younger patients (21% in regorafenib-treated patients ≥ 65 years compared to 9% in regorafenib-treated patients < 65 years).

The proportion of patients who discontinued study drug primarily due to an adverse event was similar between arms (10% and 9% in the regorafenib and placebo arms, respectively).

On May 23, 2017, Bayer informed FDA of updated safety data for some of their patients, largely from 1 investigator site in China and 1 investigator site in Taiwan that were the result of data audits performed by Bayer. Most of the new data that impacted AE reporting were Grade 1 events and affected the incidence rate of a limited number of adverse events described in labeling by 1% or less. Due to the limited scope of these changes, the label was not updated to reflect these changes.

In summary, based on these data from Study 15982, the overall safety profile was within what was expected for regorafenib in this patient population although patients ≥ 65 years appear to be at an increased risk for moderate to severe hypertension.

7.1 Methods

7.1.1 Studies/Clinical Trials Used to Evaluate Safety

The safety and efficacy analyses were primarily based on the evaluation of one trial, 15982 (PH-38451), “A randomized, double blind, placebo-controlled, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib.”

7.1.2 Categorization of Adverse Events

The severity of adverse events was documented using NCI-CTCAE version 4.0. The MedDRA 19.0 dictionary was used to code adverse event data.

Verbatim terms in the adverse event dataset were reviewed to determine whether MedDRA preferred terms were appropriately coded. Overall the coding of adverse events appeared to be adequate.

A subset of the CRFs was examined to assess the accuracy transfer of the data from the CRFs to the datasets. The data transfer was determined to be appropriate.

39 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

7.1.3 Pooling of Data Across Studies/Clinical Trials to Estimate and Compare Incidence

Not applicable.

7.2 Adequacy of Safety Assessments

7.2.1 Overall Exposure at Appropriate Doses/Durations and Demographics of Target Populations

All safety analyses were based on the safety population, which included all randomized patients who received any amount of study treatment, and were analyzed according to actual treatment received. Five patients randomized to regorafenib did not receive treatment and 1 patient randomized to placebo did not receive placebo. Thus, the safety population included the 573 randomized patients minus the above 6 patients for a total of 567 patients.

7.2.2 Explorations for Dose Response

Not applicable.

7.2.3 Special Animal and/or In Vitro Testing

Not applicable.

7.2.4 Routine Clinical Testing

Routine clinical testing and monitoring were analyzed, and the results of these analyses are described in the Laboratory and Safety Sections of this review (Sections 7.3 and 7.4).

7.2.5 Metabolic, Clearance, and Interaction Workup

Not applicable.

7.2.6 Evaluation for Potential Adverse Events for Similar Drugs in Drug Class

The safety profile of anti-VEGF agents was characterized by assessing the occurrence of hypertension, proteinuria, arterial and venous thromboembolic events, hemorrhagic events (e.g., gastrointestinal bleeding, hemoptysis, epistaxis), compromised wound healing, reversible posterior leukoencephalopathy syndrome, gastrointestinal perforation and fistula, and cardiac dysfunction. The safety of small molecule TKIs was

40 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

characterized assessing the occurrence of hypertension, proteinuria, skin reactions, thromboemboli, and diarrhea.

7 .3 Major Safety Results

The following analyses in the safety section of this review are based on adverse events that occurred between the first administration of regorafenib or placebo and 30 days after the last administration using the treatment-emergent adverse event (TEAE) flag for data set analyses. Importantly, the protocol-specified window for the safety follow up period is 30+4 days and is, thus, not identical to the population used for the main safety summary analyses in the clinical study report which used the applicant's treatment emergent adverse flag. The safety population includes the 573 randomized patients minus the 6 patients who did not receive study medication for a total of 567 patients, 193 who received placebo and 374 who received regorafenib. Table 7 summarizes the major safety results that occurred using the TEAE flag for selection. Almost all (93%) patients in the placebo arm experienced an AE compared to 100% of the patients in the regorafenib arm. The percentage of patients who experienced a Grade 3-4 AE was higher in the regorafenib arm, with 79% compared to 56% in the placebo arm. Of note, Bayer calculated the incidence of Grade 3 and 4 TEAEs differently, likely excluding from their Grade 3-4 list those patients who also had a Grade 5 event. Thus, in Bayer's Table 10-2, the Grade 3-4 numbers are 75 (38.9%) and 248 (66.3%) for the placebo and regorafenib arms, respectively (the patients who experienced a Grade 5 event were described separately). The trend remains the same using either calculation method. In contrast to AEs, the percentage of patients who experienced either an SAE or a Grade 5 AE was higher in the placebo arm compared to the regorafenib arm.

Table 7-Study 15982: Major safety results summary

no. of PL (% of N) no. of Regorafenib (% of N) N=193 N=374 Subjects who experienced an AE 179 (93) 374 (100) Subjects who experienced an AE Grade 3-4 109 (56) 295 (79) Subjects who experienced a SAE 90 (47) 166 (44) Deaths related to an AE 38 (20) 50 (13)

At the SOC level, Table 8, the most frequently affected systems of all Grades(~ 20% incidence) were Gastrointestinal disorders (78% and 59% in the regorafenib and placebo arms, respectively), General disorders and administration site conditions (70% and 55% in the regorafenib and placebo arms respectively), Skin and subcutaneous tissue disorders (66% and 31% in the regorafenib and placebo arms, respectively), Investigations (58% and 37% in the regorafenib and placebo arms, respectively), Metabolism and nutrition disorders (53% and 34% in the regorafenib and placebo arms, respectively), Respiratory, thoracic and mediastinal disorders (41% and 22% in the

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

regorafenib and placebo arms, respectively), Vascular disorders (36% and 13% in the regorafenib and placebo arms, respectively), Musculoskeletal and connective tissue disorders (35% and 28% in the regorafenib and placebo arms, respectively ), and Infections and infestations (31 % and 18% in the regorafenib and placebo arms, respectively).

Table 8- Study 15982: AEs by SOC no. of PL (% of N) no. of Regorafenib (% of N) soc N=193 N=374 All Grades Grades 3-5 All Grades Grades 3-5 Gastrointestinal disorders 114 (59) 29 (15) 290 (78) 65 (17) General disorders and administration site conditions 106(55) 31 (16) 260 (70) 71 (19) Skin and subcutaneous tissue disorders 59 (31) 2 (1) 245 (66) 51 (14) Investigations 72 (37) 40 (21) 217(58) 117 (31) Metabolism and nutrition disorders 66 (34) 22 (11) 199 (53) 75 (20) Respiratoty, thoracic and media.stinal disorders 43 (22) 11 (6) 153 (41) 17 (5) Vascular disorders 26 (13) 11 (6) 134 (36) 60 (16) Musculoskeletal and connective tissue disorders 55 (28) 8 (4) 132 (35) 17 (5) Infections and infestations 35 (18) 11 (6) 117 (31) 30 (8) Nervous system disorders 49 (25) 8 (4) 91 (24) 22 (6) Blood and lymphatic system disorders 30 (16) 12 (6) 65 (17) 25 (7) Renal and urinaty disorders 18 (9) 4 (2) 63 (17) 13 (3) Hepatobilia1y disorders 31 (16) 26 (13) 55 (15) 32 (9) Psychiatric disorders 17 (9) 2 (1) 45 (12) 4 (1) Endocrine disorders 0 (0) 0 (0) 29 (8) 0 (0) Cardiac disorders 9 (5) 1 (1) 27 (7) 7 (2) Injmy, poisoning and procedural complications 14 (7) 3 (2) 24 (6) 4 (1) Ear at1d labyrinth disorders 5 (3) 0 (0) 19 (5) 0 (0) Reproductive system and breast disorders 10 (5) 2 (1) 18 (5) 0 (0) Neoplasms benign, malignant and unspecified (incl cysts and polyps) 8 (4) 2 (1) 15 (4) 10 (3) Eye disorders 5 (3) 0 (0) 9 (2) 1 (0) Surgical and medical procedures 0 (0) 0 (0) 2 (1) 2 (1) Iinmm1e system disorders 0 (0) 0 (0) 1 (0) 0 (0)

At the preferred term (PT) level,

Table 9, the most frequently reported adverse events of all Grades (2: 20% incidence) were palmar-plantar erythrodysaesthesia syndrome (PPES) (51% and 7% in the

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

regorafenib and placebo arms, respectively), diarrhea (41 % and 15% in the regorafenib and placebo arms, respectively), hypertension (31% and 6% in the regorafenib and placebo arms, respectively), decreased appetite (30% and 14% in the regorafenib and placebo arms, respectively), fatigue 28% and 24% in the regorafenib and placebo arms, respectively), increased aspartate aminotransferase (AST) (25% and 20% in the regorafenib and placebo arms, respectively), increased blood bilirubin (24% and 16% in the regorafenib and placebo arms, respectively), abdominal pain (21 % and 16% in the regorafenib and placebo arms, respectively), and pyrexia (20% and 7% in the regorafenib and placebo arms, respectively).

The complete list of AEs (all grades) by PT can be found in Table 31. FDA's analysis concurs with Bayer's reported incidence rates of TEAEs in their clinical study report.

Table 9- Study 15982: AEs by PT with~ 5% incidence in regorafenib arm and with a~ 3% difference between arms

no. of o/o of no. of PL o/o of PL PT Regora.fenib Regorafenib N=193 N = 374 Palmar-plantar erythrodysaesthesia 13 7 192 51 syndrome (PPES) Dian·hoea 29 15 153 41 H ype1tension 12 6 115 31 Decreased appetite 27 14 114 30 Fatigue 47 24 106 28 Aspattate aminotransferase (AST) 38 20 92 25 increased Blood bilirubin increased 31 16 91 24 Abdominal pain 30 16 79 21 Pyrexi a 13 7 74 20 Dysphonia 3 2 66 18 Constipation 21 11 65 17 Nausea 26 13 64 17 Asthenia 18 9 56 15 Alanine aininotransferase increased 21 11 54 14 Hvooalbtuninaemia 14 7 52 14 Anaemia 21 11 51 14 Weight decreased 8 4 50 13 Abdominal pain unoer 17 9 47 13 Vomiting 13 7 47 13 Back pain 17 9 45 12 Cough 13 7 41 11 Muscle spasms 4 2 38 10 Hvooohosphatae1nia 4 2 36 10 Platelet count decreased 2 1 34 9 Proteinuria 2 1 32 9 Stomatitis 4 2 31 8 Lipase increased 6 3 27 7 Alopecia 5 3 26 7

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

no. of %of no. of PL % ofPL PT Regorafenib Regorafenib N = 193 N=374 H wokalaemia 5 3 26 7 Pain in extremitv 6 3 26 7 H ypothvroidism 0 0 24 6 Malaise 5 3 22 6 H yponatraemia 6 3 21 6 White blood cell count decreased 2 1 17 5

When grouped by high level term (HLT), Table 10, the most frequent AEs (;::: 20%) were skin and subcutaneous conditions, asthenic conditions, diarrhea, liver function analyses, gastrointestinal and abdominal pain, appetite disorders, vascular hypertensive disorders, musculoskeletal and connective tissue pain and discomfort, nausea and vomiting symptoms, upper respiratory tract signs and symptoms, and febrile disorders.

Table 10- Study 15982: AEs by HLT (incidence;::: 5%) %of no.PL %ofPL no. Regorafenib BLT Regorafenib N=193 N=374 Skin and subcutaneous conditions NEC 13 7 192 51 Asthenic conditions 67 35 174 47 Diarrhoea (excl infective) 29 15 154 41 Liver function analyses 57 30 141 38 Gastrointestinal and abdominal pains ( excl oral and throat) 47 24 121 32 Appetite disorders 28 15 115 31 Vascular hypertensive disorders NEC 12 6 115 31 Musculoskeletal and C·01ltlective tissue pain and discomfort 35 18 91 24 Nausea and vomiting symptoms 32 17 83 22 Upper respirato1y tract signs and symptoms 8 4 77 21 Febrile disorders 13 7 74 20 Gastrointestinal atonic and hypomotility disorders NEC 24 12 66 18 Oedema NEC 26 13 62 17 Peritoneal and retroperitoneal disorders 31 16 58 16 Physical examination procedmes and organ system status 9 5 56 15 General signs and symptoms NEC 35 18 55 15 Protein metabolism disorders NEC 15 8 54 14 Anaemias NEC 21 11 51 14

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

o/o of no.PL o/o of PL no. Regorafenib HLT Regorafenib N=193 N=374 Coughing and associated symptoms 17 9 45 12 Muscle related signs and symptoms NEC 4 2 39 10 Urinary abnonnalities 6 3 39 10 Potassium imbalance 12 6 37 10 Phosphorns metabolism disorders 4 2 36 10 Platelet analyses 2 1 36 10 Stomatitis and ulceration 6 3 36 10 Breatl1ing abnonnalities 17 9 32 9 White blood cell analyses 5 3 32 9 Tissue enzyme analyses NEC 11 6 30 8 Digestive enzymes 6 3 29 8 Lower respirato1y tract and lung infections 5 3 28 7 Upper respiratory tract infections 5 3 27 7 Alopecias 5 3 26 7 Rashes, emptions and exanthems NEC 15 8 26 7 Headaches NEC 12 6 24 6 Pain and discomfort NEC 9 5 24 6 Thyroid hypofunction disorders 0 0 24 6 Cholestasis and jaundice 8 4 23 6 Flatulence, bloating and distension 12 6 22 6 Oral diyness and saliva altered 9 5 22 6 Sodium imbalance 7 4 21 6 Pmritus NEC 15 8 20 5 Dennal and epidermal conditions NEC 10 5 18 5 Nasal disorders NEC 2 1 18 5

7 .3.1 Deaths

Eighty eight deaths occurred during the TEAE (30 day) period, 50 (13%) in the regorafenib arm and 38 (20%) in the placebo arm. Table 11 details the causes of death by PT in each arm during the TEAE period. The two arms are generally balanced with respect to the causes and, based on a focused review of narratives and brief case summaries, many of these deaths appeared to occur in the setting of clinical progression of disease.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table 11 - Study 15982: Causes of Grade 5 events by PT no.of no. of PL Regol'afenib

Deaths 38 50 Causes by PT 39* 50 Acute hepatic failure 0 1 Ascites I 2 Blood pressure decreased 0 1 Bronchial obstmction 0 1 Cardiac atl'est I 0 Craniocerebral iniurv 0 1 Death 0 1 Duodenal perforation 0 1 Dvspnoea I 2 Encephalopathy I 0 General physical health 16 23 deterioration Haemorrhage intra.crruiial 0 1 Hepatic encephalopathy I 1 Hepatic faihu·e 5 3 Hepatic h.aemorrhage 2 0 Hepatorenal syndrome I 1 Hypovolaemic shock 0 1 Intra-abdominal haemorrhage I 0 Lung infoction 0 1 Me1i.ingorrhagia 0 1 Multiple organ dysfimction I 0 svndrome Mvocardial infarction 0 1 Oesophageal varices haemorrhage I 0 Peritonitis bacterial 0 1 Plew·al effusion I 0 Pnetunonia 0 1 Respiratorv failure 3 1 Sepsis 0 1 Septic shock 0 1 Shock haemorrhagic 0 2 Tumour haemoIThage I 0 Unner gastrointestinal haemoll'hage 2 0 Note: 39 different Grade 5 events were attributed to 38 individual patients in the placebo arm. One patient in the placebo aim had two Grade 5 events attributed, one each in Gastrointestinal disorders (ascites) and Nervous system disorders (hepatic encephalopathy).

Of these 88 deaths, the SOC groups (Table 12) that differed the most between the two arms and for which the regorafenib arm had a higher incidence were Infections and infestations (1.3% and 0% in the regorafenib and placebo arms, respectively) and Vascular disorders (0.8% and 0% in the regorafenib and placebo arms, respectively). Review of these narratives did not reveal any emerging patterns. One patient died from

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

meningorrhagia in the setting of regorafenib and therapeutic low molecular weight heparin (LMWH) but in the absence of CNS lesions. This event is consistent with prior regorafenib adverse event data.

Table 12- Causes of Grade 5 events by SOC no. of no. of PL Regorafenib (o/o ofN) soc (o/o ofN) N = 193 N = 374

Cardiac disorders 1 (0.5) 1 (0.3) Gastrointestinal disorders 5 (2.6) 3 (0.8) General disorders and 17 (8.8) 24 (6.4) administration site conditions Hepatobilia1y disorders 8 (4.1) 5 (1.3) Infections and infestations 0 (0) 5 (1.3) Injwy, poisoning and procedmal 0 (0) 1 (0.3) complications Investigations 0 (0) 1 (0.3) Neoplasms benign, malignant and 1 (0.5) 0 (0) unspecified (incl cysts and polyps) Nervous system disorders 2 (1 .0) 3 (0.8) Respiratory, thoracic and 5 (2.6) 4 (1.1) mediastinal disorders Vascular disorders 0 (0) 3 (0.8) Note: One patient in the placebo ann had two Grade 5 events attributed, one each in Gastrointestinal disorders (ascites) and Nervous system disorders (hepatic encephalopathy).

Of the nine deaths attributed by investigators to study drug, two occurred in the placebo arm (both with PT of hepatic failure) and seven occurred in the regorafenib arm (with PTs of duodenal perforation, meningorrhagia, shock hemorrhagic, hepatic encephalopathy, death (NOS), myocardial infarction, and general physical health deterioration). Review of these narratives did not reveal unexpected patterns for this drug class in this population of patients.

7 .3.2 Nonfatal Serious Adverse Events

During the 30-day TEAE period, the incidence of SA Es was 44 % in the regorafenib arm and 47% in the placebo arm. Table 13 lists the incidences of non-fatal SAEs by PT and Table 14 lists the incidence of Grade 3 and 4 SAEs by SOC that occurred in the two arms. These incidence rates are comparable between the two arms.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table 13- Study 15982: Non-fatal (Grades 1-4) SAEs (incidence ;:;;1%) by PT no. of Regorafenib no. of PL(% ofN) PT (% ofN) N=193 N=374 General physical health deterioration 13 (7) 29 (8) Ascites 6 (3) 9 (2) Hepatic failure 8 (4) 9 (2) Hepatic encephalopathy 2 (I) 7 (2) Back pain 2 (I) 6 (2) Pneumonia 1(1) 6 (2) Dyspnoea 2 (I) 5 ( 1) Pvrexia 1(1) 5 ( I) Oesophageal varices haemoffhage 1(1) 4 ( 1) Plew-al effusion 1(1) 4 ( I) Twnourpain 0 (0) 4 ( 1) Acute coronarv svndrome 0 (0) 3 ( I) Anaemia 1(1) 3 ( 1) Asthenia 0 (0) 3 ( I) Dehydration 0 (0) 3 ( 1) Diarrhoea 0 (0) 3 ( I) Ence.phalooathy 3 (2) 3 ( 1) Haemontvsis 2 (I) 3 ( I) Pancreatitis 0 (0) 3 ( 1) Shock haemon-hagic 0 (0) 3 ( I) Unner gastrointestinal haemonhage 3 (2) 3 ( 1) Abdominal infection 0 (0) 2 ( I) Abdominal pain 4 (2) 2 ( 1) Abdominal pain lower 0 (0) 2 ( I) Atrial fibrillation 0 (0) 2 ( 1) Fati~me 0 (0) 2 ( I) Hepatic ciffhosis 0 (0) 2 ( 1) Hepatic function abnonnal 3 (2) 2 ( I) Hepatic haem01Thage 2 (I) 2 ( 1) Hypoglycaemia 0 (0) 2 ( I) H voonatrae1nia 0 (0) 2 ( 1) Jaundice cholestatic 1(1) 2 ( I) Liver abscess 2 (I) 2 ( 1) Lung infe.ct.ion 0 (0) 2 ( I) Pneumonitis 0 (0) 2 ( 1) Renal failure 1(1) 2 ( I) Seizure 0 (0) 2 ( 1) Sepsis 0 (0) 2 ( I)

Table 14- Study 15982: Grade 3-4 SAEs by SOC no. of Regorafenib no. of PL(% ofN) soc (% ofN) N=193 N=374 General disorders and administration 14 (7) 31 (8)

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

no. of Regorafenib no. of PL (% of N) soc (% ofN) N=193 N=374 site conditions Gastrointestinal disorders 17 (9) 27 (7) Infections and infestations 5 (3) 25 (7) Hepatobiliruy disorders 20 (10) 21 (6) Nervous system disorders 4 (2) 18 (5) Respiratory, thoracic and mediastinal disorders 8 (4) 14 (4) Musculoskeletal and connective tissue disorders 4 (2) 10 (3) Metabolism and nutrition disorders 4 (2) 8 (2) Neoplasms benign, malignant and unspecified (incl cysts and polyps) 1 (1) 8 (2) Cardiac disorders 1 (1) 6 (2) Vascular disorders 2 (1) 6 (2) Blood and lymphatic system disorders 2 (1) 4 (1) Injury, poisoning and procedural complications 3 (2) 4 (1) Investigations 2 (1) 4 (1) Renal and urinaty disorders 2 (1) 4 (1) Skin and subcutaneous tissue disorders 0 (0) 2 (1) Eye disorders 0 (0) 1 (0) Psychiatric disorders 0 (0) 1 (0) Reproductive system and breast disorders 1 (1) 0 (0)

7.3.3 Dropouts and/or Discontinuations

Dose reductions and interruptions occurred more often in the regorafenib arm compared to the placebo arm wh ile both arms had similar incidence rates of discontinuations (Table 15). Of note, a patient could have more than one dose reduction or interruption. The most frequent AEs leading to study drug withdrawal in the regorafenib arm were general physical health deterioration/abnormal general physical health condition (6 events), hepatic failure (5 events), increased AST (3 events), hepatic encephalopathy (3 events), increased blood bilirubin/increased conjugated bilirubin/hyperbilirubinemia (3 events), fatigue/asthenia (3 events), and PPES (2 events).

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table 15- Study 15982: Study drug dose reduced, interrupted, or withdrawn no. of no. of PL o/o of o/o of PL Regorafenib Event patients Regorafenib N = 193 patients N = 374

Total reduced, interrupted, 0 1· 46 24 225 60 withdrawn Reduced 9 5 105 28 Intenupted 29 15 172 46 Withdrawn (discontinued) 18 9 37 10

This analysis differs from that of Bayer with their analysis concluding that dose interruptions and dose reductions occurred in ~ % and 48% of regorafenib-treated patients, respectively. This issue (and reasons for discrepancies) is currently under review as part of labeling.

7.3.4 Significant Adverse Events- Non-Fatal Grade 3-4 AEs

This section focuses on Grade 3-4 AEs. Adverse events that are known to be related to VEGF inhibition will be further discussed in Section 7.3.5 Submission Specific Primary Safety Concerns. Table 16 summarizes the incidence of all Grade 3-4 AEs by SOC.

Table 16- Study 15982: Grade 3-4 AEs by SOC

no. ofGr3-4 o/o of PL no. of Gr 3-4 o/o of soc PL N = 193 Regorafenib Regorafenib N = 374 Investigations 40 21 116 31 Metabolism and nutrition disorders 22 11 75 20 Gastrointestinal disorders 24 12 62 17 Vascular disorders 11 6 57 15 Skin and subcutaneous tissue 2 l 51 14 disorders General disorders and administration 14 7 47 13 site conditions Hepatobiliary disorders 18 9 27 7 Blood and lymphatic system 12 6 25 7 disorders Infections and infestations 11 6 25 7 Nervous system disorders 6 3 19 5 Musculoskeletal and connective 8 4 17 5 tissue disorders Renal and urinary disorders 4 2 13 3 Respiratory, thoracic and mediastinal 6 3 13 3

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

no. of Gr 3-4 o/o of PL no. of Gr 3-4 o/o of soc PL N = 193 Regorafenib Regol'llfenib N=374 disorders Neoplasms benign, malignant and 1 1 10 3 unspecified (incl cysts and polyps) Cardiac disorders 0 0 6 2 Psychiatric disorders 2 1 4 1 Injury, poisoning and procedural 3 2 3 1 complications Surgical and medical procedures 0 0 2 1 Eye disorders 0 0 1 0 Reproductive system and breast 2 1 0 0 disorders

Table 17 summarizes the most frequent Grade 3-4 AEs by PT, regardless of outcome, which occurred with an incidence rate of 2% or more. Of these, hypertension (15% and 5% in the regorafenib and placebo arms, respectively), PPES (12% and 1% in the regorafenib and placebo arms, respectively), hypophosphatemia (8% and 2% in the regorafenib and placebo arms, respectively), and increased lipase (7% and 2% in the regorafenib and placebo arms, respectively) were the most common. Increased AST and increased blood bilirubin were also among the most common Grade 3-4 AEs but were not increased in the regorafenib arm compared to the placebo arm.

Table 17 - Study 15982: Grade 3-4 AEs by PT (incidence ;::: 2%)

no. of Gr 3-4 no. of Gr 3-4 o/o of o/o of PL PT PL Regol'llfenib Regol'llfenib N=193 N=193 N=374 N=374 H voe1tension 9 5 55 15 Palmar-plantar erythrodysaesthesia 1 1 46 12 svndrome Aspartate aminotransferase 22 11 41 11 increased Hvooohosphataemia 3 2 31 8 Blood bilimbin increased 18 9 28 7 Lipase increased 3 2 25 7 Fati21.1e 7 4 22 6 Ascites 11 6 16 4 Anaemia 11 6 15 4 General ohvsical health deterioration 9 5 15 4 H voonatraemia 6 3 15 4 Asthenia 2 1 14 4 DiaiThoea 0 0 12 3 Gamma-glutamyltransferase 5 3 12 3 increased

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

no. of Gr 3-4 no. of Gr 3-4 %of % ofPL PT PL Regorafenib Regorafenib N=193 N=193 N=374 N=374 Alanine aminotransferase increased 5 3 11 3 Abdominal pain 5 3 10 3 Decreased annetite 3 2 10 3 Platelet count decreased 0 0 10 3 H vookalaemia 2 1 9 2 Back pain 2 1 8 2 Blood alkaline phosphatase 4 2 7 2 increased H yperbilirubinaemia 3 2 7 2 Proteinuria 1 1 7 2 Weight decreased 0 0 7 2 Amylase increased 0 0 6 2 Hepatic encephalopathy 1 1 6 2 Hepatic faihu·e 4 2 6 2 H ypoalbuminaemia 1 1 6 2

Adverse events with an increased incidence of 2% or more in the regorafenib arm compared to the placebo arm were hypertension, palmar-plantar erythrodysaesthesia syndrome (PPES), hypophosphatemia, increased lipase, fatigue, asthenia, diarrhea, decreased platelet count, decreased weight, and increased amylase.

As discussed above, FDA's analysis of the overall incidence of Grade 3-4 TEAEs differed from that of Bayer though the trends were the same.

Investigations As an SOC, Investigations were the most frequently observed Grade 3-4 TEAEs in Study 15982 and Table 18 summarizes the individual PTs under this SOC. Events with a higher incidence in the regorafenib arm were increased lipase, decreased platelet count, decreased weight, increased amylase, decreased white blood cell count, decreased hemoglobin, and decreased lymphocytes. A higher incidence rate of Grade 3-4 blood bilirubin levels was observed in the placebo arm.

Table 18- Study 15982: Grade 3-4 Investigations AEs by PT (and incidence 2::1% in regorafenib arm) no. of Gr 3-4 no. of Gr 3-4 %of %ofPL PT PL Regorafenib Regorafenib N=193 N=193 N=374 N=374 Aspartate aminotransferase increased 22 11 41 11 Blood bilirubin increased 18 9 28 7 Lipase increased 3 2 25 7 Gamma-glutamyltransferase increased 5 3 12 3

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Alanine aminotransferase increased 5 3 11 3 Platelet count decreased 0 0 10 3 Blood alkaline phosphatase increased 4 2 7 2 Weight decreased 0 0 7 2 Amylase increased 0 0 6 2 Bilirubin conjugated increased 1 1 5 1 Neutrophil cotmt decreased 1 1 4 1 White blood cell cotmt decreased 0 0 4 1 Haemoglobin decreased 0 0 3 1 Lymphocyte count decreased 0 0 3 1

Metabolism and nutrition disorders Metabolism and nutrition disorders was the next most frequent SOC in terms of Grade 3-4 TEAEs observed and the individual PTs are summarized in Table 19. Hypophosphatemia occurred at an increased incidence in the regorafenib arm.

Table 19- Study 15982: Grade 3-4 Metabolism and Nutrition Disorder A Es by PT (and incidence ;::1 % in regorafenib arm)

no. of Gr 3-4 no. of Gr 3-4 o/o of o/o of PL PT PL Regorafenib Regorafenib N= 193 N=193 N=374 N=374 Hypophosphataemia 3 2 31 8 H yponatraemia 6 3 15 4 Decreased appetite 3 2 10 3 Hypokalaemia 2 1 9 2 Hypoalbuminaemia 1 1 6 2 Dehydration 0 0 5 1 Hyperglycaemia 4 2 5 1 H yperkalaemia 2 1 4 1 Hypoglycaemia 0 0 4 1 Hyperuricaemia 0 0 2 1 Hypocalcaemia 0 0 2 1

Gastrointestinal disorders Gastrointestinal disorders are of particular interest in this HCC population and Table 20 summarizes the Grade 3-4 events in this SOC. Diarrhea occurred at an incidence of 3% in the regorafenib arm compared to 0% in the placebo arm while the other PTs were similar between the two arms.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table 20- Study 15982: Grade 3-4 Gastrointestinal Disorder AEs by PT (and incidence ;::1% in regorafenib arm)

no. of Gr 3-4 no. of Gr 3-4 o/o of o/o of PL PT PL Regorafenib Regorafenib N=193 N=193 N=374 N=374 Ascites 11 6 18 5 DiaIThoea 0 0 12 3 Abdominal pain 5 3 10 3 Abdominal pain lower 0 0 3 1 Oesophageal varices haemoIThage 1 1 3 1 Stomati tis 0 0 3 1 Upper gastrointestinal haemo1Thage 3 2 3 1 Vomiting 1 1 3 1 Abdominal pain upper 2 1 2 1 Dysphagia 0 0 2 1 Gastritis 1 1 2 1 Nausea 0 0 2 1

Vascular disorders Vascular disorders are also of particular interest in the HCC population and in patients taking VEGF-inhibitors and are summarized in Table 21 . The incidence rate of hypertension was higher in the regorafenib arm as was the incidence rate of Grade 3 hypertension.

Table 21- Study 15982: Grade 3-4 Vascular Disorder AEs by PT (and incidence ;::1 % in regorafenib arm)

no. of Gr 3-4* no. of Gr 3-4 o/o of o/o of PL PT PL Regorafenib Regorafenib N=193 N=193 N=374 N=374 Hypertension 9 5 55 15 Shock haemo1Thagic 0 0 3 1

Skin and subcutaneous tissue disorders The increased incidence of Grade 3-4 skin and subcutaneous tissue disorders (in Table 16) in the regorafenib arm compared to the placebo arm is largely due to the increased incidence of PPES as shown in Table 22. No Grade 4 Skin and subcutaneous tissue disorder adverse events were reported in either arm.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table 22- Study 15982: Grade 3-4 Skin and subcutaneous tissue disorder AEs by PT (and incidence <::1 % in regorafenib arm)

no. of Gr 3-4 no. of Gr 3-4 o/o of o/o of PL PT PL Regorafenib Regorafenib N= 193 N=193 N=374 N=374 Pal:tnar-plantar erythrodysaesthesia syndrome 1 1 46 12 Skin ulcer 0 0 2 1

General disorders and administration site conditions While general physical health deterioration events of all Grades occurred at an increased incidence in the placebo arm compared to the regorafenib arm (14% compared to 12%, Table 31 ), Grade 3-4 general physical health deterioration events occurred at a similar incidence between the two arms (Table 23) while Grade 3-4 events of asthenia and fatigue occurred at an increased incidence in the regorafenib arm compared to the placebo arm.

Table 23- Study 15982: Grade 3-4 General disorders and administration site condition AEs by PT (and incidence <::1 % in regorafenib arm)

no. of Gr 3-4 no. of Gr 3-4 o/o of o/o of PL PT PL Regorafenib Regorafenib N=193 N=193 N=374 N=374 General physical health deterioration 14 7 28 7 Fatigue 7 4 22 6 Asthenia 2 1 14 4

Hepatobiliary disorders Although not increased in incidence in the regorafenib arm, because of their particular relevance to this population, the Grade 3-4 hepatobiliary disorder PTs are summarized in Table 24. There was an increased incidence of hepatic failure in the placebo arm compared to the regorafenib arm which likely reflects the increased incidence of progression of disease in the placebo arm.

Table 24- Study 15982: Grade 3-4 Hepatobiliary disorder AEs by PT

no. of Gr 3-4 no. of Gr 3-4 o/o of o/o of PL PT PL Regorafenib Regorafenib N=193 N=193 N=374 N=374 Hepatic faihu·e 8 4 9 2 H yperbilirubinaemia 3 2 7 2 Hepatic function abno1mal 4 2 2 1

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Hepatic haemotThage 2 1 2 1 Hepatorenal syndrome 2 1 2 1 Jaundice 3 2 2 1 Jam1dice cholestatic 1 1 2 1 Acute hepatic failure 0 0 1 0 Bile duct stone 0 0 1 0 Cholangitis 1 1 1 0 Cholecystitis 0 0 1 0 Gallbladder obstmction 0 0 1 0 Hepatic cin-hosis 1 1 1 0 Hepatic ischaemia 0 0 1 0 Hepatitis acute 0 0 1 0 Hepatobiliary disease 0 0 1 0 Po1ial vein thrombosis 1 1 1 0 Bile duct stenosis 4 2 0 0 Cholestasis 1 1 0 0

Infections and infestations Grade 3-4 infections and infestation PTs are summarized in Table 25. Based on HlTs for AEs of all grades (Table 10), lower respiratory tract and lung infections and upper respiratory tract infections occurred at an increased incidence in the regorafenib arm compared to the placebo arm and Table 25 reflects this trend as well. To address the increased incidence rate of infections, the applicant proposed a Warning describing the overall increased rate of infections in the pooled randomized trials (HCC trial plus previously reviewed colorectal cancer and GIST trials).

Table 25- Study 15982: Grade 3-4 Infections and infestation AEs by PT (and incidence <::1% in regorafenib arm)

no. of Gr 3-4 no. of Gr 3-4 o/o of o/o of PL PT PL Regorafenib Regorafenib N=193 N=193 N=374 N=374 Pneumonia 1 1 6 2 Abdominal infection 2 1 3 1 Sepsis 0 0 3 1 Lung infection 1 1 2 1 Peritonitis bacterial 1 1 2 1 Urinary tract infection 0 0 2 1

7.3.5 Submission Specific Primary Safety Concerns

Safety issues with particular relevance to this anti-VEGF TKI, were hypertension, proteinuria, gastrointestinal fistulas and perforation, wound healing difficulty,

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

thromboemboli, and cardiac events. Adverse event, laboratory, and vital sign data were examined and do not reveal any new safety findings with the exception of the magnitude of the incidence of Grade 3-5 hypertension in patients older than 65 years old (Section 7.5.3, below).

7 .4 Supportive Safety Results

7 .4.1 Common Adverse Events

Table 26 summarizes Grade 1-2 events.

Table 26- Study 15982: Grade 1-2 AEs by PT (incidence> 5%) no. of Gr 1-2 no. of Gr 1-2 o/oof o/oofPL PT PL Regora.fenib Regorafenib N= 193 N=193 N=374 N=374 Palmar-plantar erythrodysaesthesia 13 7 187 50 syndrome DiaIThoea 29 15 151 40 Decreased appetite 25 13 111 30 Fatigue 42 22 102 27 H ypettension 9 5 94 25 Blood bilimbin increased 24 12 79 21 Abdominal pain 26 13 75 20 Aspattate aminotransferase 30 16 75 20 increased Pyrexi a 13 7 74 20 Dysphonia 3 2 66 18 Constipation 20 10 64 17 Nausea 26 13 64 17 Oedema peripheral 26 13 56 15 Asthenia 18 9 53 14 Hypoalbuminaemia 13 7 50 13 Weight decreased 8 4 50 13 Ascites 27 14 49 13 Alanine aminotransferase increased 19 10 48 13 Abdominal pain upper 16 8 47 13 Anaemia 17 9 47 13 Vomiting 12 6 46 12 Back pain 15 8 43 11 Cough 13 7 41 11 Muscle spasms 4 2 38 10 Platelet count decreased 2 1 32 9 Stomatitis 4 2 31 8

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Proteinuria 2 1 29 8 Alopecia 5 3 26 7 Dyspnoea 12 6 26 7 Pain in extremity 6 3 26 7 Hypothyroidism 0 0 24 6 Headache 12 6 23 6 Insomnia 8 4 23 6 Malaise 5 3 22 6 Dry mouth 9 5 21 6 Hypokalaemia 4 2 21 6

7.4.2 Laboratory Findings

Analysis of selected hematologic, metabolic, gastrointestinal, renal, and endocrine laboratory data are displayed in Table 27 and are consistent with data provided in the submission. Clinically relevant findings in the regorafenib arm compared to the placebo arm were reported for neutropenia (Grade 3), thrombocytopenia, hypocalcemia, hypokalemia, hypophosphatemia, hyperbilirubinemia, increased alkaline phosphatase, increased ALT, increased AST, increased INR, increased lipase, proteinuria, and hypoglycemia. These findings are consistent with this class of drugs and with prior experience with regorafenib. Although the incidence of the hypothyroidism AE is increased in the regorafenib arm compared to placebo (6% versus 0%, Table 9), there was less of a difference in the laboratory parameter of decreased free T4 between the two arms (5% versus 3%). The incidence of increased creatinine was lower in the regorafenib arm compared to the placebo arm (77% versus 86%).

58 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table 27 - Study 15982: Laboratory data no. of PL (% of N) no. of Regorafenib (% of N) Table 27 N= 193* N=374* All Grades Grade3 Grade 4 All Grades Grade3 Grade4 Anemia 134 (71) 9 (5) 0 (0) 266 (72) 22 (6) 0 (0) Lymphopenia 110 (59) 21(11) 1 (0.5) 249 (68) 57 (16) 7 (2) Neutropenia 28 (15) 1 (0.5) 1 (0.5) 50 (14) 11 (3) 0 (0) Thrombocytopenia 94 (50) 0 (0) 0 (0) 231 (63) 17 (5) 3 ( 0.8) H ypocalcemia 19 (10) 0 (0) 0 (0) 86 (23) 1 (0.3) 0 (0) Hypokalemia 17 (9) 4 (2) 0 (0) 113 (31) 14 (4) 2 (0.5) Hypophosphatemia 59 (31) 13 (7) 0 (0) 259 (70) 119 (32) 6 (2) H yperbilimbinemia 104 (54) 21 (11) 9 (5) 290 (78) 48 (13) 11 (3) Increased alkaline 137 (73) 19 (10) 1 (0.5) 290 (79) 35 (10) 0 (0) phosphatase Increased ALT 112 (59) 9 (5) 0 (0) 261 (70) 21 (6) 2 (0.5) Increased amylase 35 (19) 4 (2) 1 (0.5) 84 (23) 9 (2) 1 (0.3) Increased AST 161 (84) 33 (17) 5 (3) 344 (93) 60 (16) 6 (2) Increased GGT 168 (89) 76 (40) 5 (3) 325 (88) 128 (35) 13 (4) Increased INR 51 (35) 3(2) 0 (0) 125 (44) 2 (0.7) 0 (0) Increased lipase 50 (27) 14 (8) 2 (1) 148 (41) 41 (11) 11 (3) Increased creatinine 161 (86) 4 (2) 1 (0.5) 282 (77) 7 (2) 0 (0) Proteinuria 58 (37) 5 (3) 0 (0) 152 (51) 50 (17) 0 (0) Decreased Free T4 6 (3) - - 19 (5) - - Hypoglycemia 15 (8) 1 (0.5) 1 (0.5) 61 (17) 5 (1) 1 (0.3) all grades = Gr 1-5 * = Based on patients with post-baseline laboratory samples.

7 .4.3 Vital Signs

In addition to reviewing hypertension as an AE, the vital sign data set was reviewed for elevations in blood pressure. Grade 2 or higher hypertension was reported in 62% of patients in the regorafenib arm compared to 40% of the patients in the placebo arm and Grade 3 or higher hypertension was reported in 21% of patients in the regorafenib arm compared to 9% of patients in the placebo arm. These results are consistent with the mechanism of action of regorafenib and with its reported adverse event profile.

7 .4.4 Electrocard iograms (ECGs)

No ECG monitoring data were submitted.

7 .4.5 Special Safety Studies/Clinical Trials

Not applicable.

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

7.4.6 Immunogenicity

Not applicable.

7.5 Other Safety Explorations

7.5.1 Dose Dependency for Adverse Events

The mean daily dose of regorafenib administered during this study was 144 mg and the median dose was 159 mg. The dose dependency for adverse events was not submitted in this application as all patients received the same starting dose in the clinical trial.

7.5.2 Time Dependency for Adverse Events

Regarding the development of PPES, 76% of regorafenib-treated patients who developed PPES, developed it during the first cycle of treatment which is consistent with the findings from patients with CRC and GIST who received regorafenib. Similarly, most of the patients treated with regorafenib who develop hypertension developed it during the first cycle of treatment.

7.5.3 Drug-Demographic Interactions

Race Table 28 summarizes selected laboratory assessments based on race. The incidence of PPES was higher in Asian regorafenib-treated patients compared to white regorafenib- treated patients (14% versus 8%) and this finding is consistent with the data reported in trials of regorafenib in patients with CRC and GIST. The incidence of increased ALT was also higher in Asian patients compared to white patients though the overall incidence was only 5%. In white patients, there was a higher incidence of increased blood bilirubin in the regorafenib-treated arm compared to the placebo arm. Due to the small number of patients, conclusions regarding Black/African American patients cannot be made. Additionally, for 20% and 23% of patients in the regorafenib and placebo arms, respectively, data on race was not reported in accordance with laws governing study sites in particular countries. In part, because of this lack of reporting, conclusions regarding the differences in ALT and bilirubin levels described above also cannot be considered conclusive.

60 Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Table 28- Study 15982: Race/Ethnicity and laboratory assessments PT o/o of Gr 3-4 no. of Gr 3-4 o/o of no. of Gr 3-4 PL Asian PL Regorafenib Regorafenib N=78 N=155 Alanine aminotransferase 1 1 7 5 increased Aspartate aminotransferase 13 17 24 15 increased Blood bilirubin increased 11 14 10 6 Hypettension 2 3 13 8 Hypophosphataemia 2 3 13 8 Palmar-plantar ecythrodysaesthesia 0 0 21 14 syndrome (PPES) no. of Gr 3-4 o/o of Black/Afr ican no. of Gr 3-4 PL o/o of PL Regorafenib Regorafenib N=2 American N=6 Alanine aminotransferase 0 0 0 0 increased Aspartate aminotransferase 0 0 1 17 increased Blood bilimbin increased 0 0 0 0 Hypertension 0 0 0 0 Hypophosphataemia 0 0 0 0 Palmar-plantar erythrodysaesthesia 0 0 0 0 syndrome (PPES) no. of Gr 3-4 o/o of no. of Gr 3-4 PL o/o of PL White Regorafenib Regorafenib N=68 N=135 Alanine aminotransferase 3 4 3 2 increased Aspartate aminotransferase 5 7 8 6 increased Blood bilirubin increased 3 4 13 10 Hype1tension 5 7 23 17 Hypophosphataemia 1 1 9 7 Palmar-plantar erythrodysaesthesia 1 1 11 8 syndrome (PPES)

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

Age Table 29 shows the incidence of Grades 1-5 and Grades 3-5 selected adverse events in the regorafenib and placebo arms based on age < 65 years or~ 65 years. Of the 193 patients in the placebo arm, 115 were< 65 years (60%) and 78 were~ 65 years (40%). Of the 374 patients in the regorafenib arm, 195 were< 65 years (52%) and 179 were~ 65 years (48%). Among patients~ 65 years, the incidence of Grades 3-5 hypertension and hypophosphatemia appeared higher compared to younger patients.

Table 29- Study 15982: Age and laboratory assessments Placebo Reeorafenib age < 65 years age 2'. 65 yeru:s age< 65 years age 2'. 65 years N = 115 N = 78 N= 195 N = 179

% % % % % % Grades % Grades %Grades Grades PT Grades Grades Grades Grades 1-5 3-5 1-5 3-5 3-5 1-5 3-5 1-5

Hypettension 5 3 8 6 26 9 36 21 Hypokalaemia 3 2 1 0 7 1 7 4 H yponatraemia 4 4 1 1 6 3 6 5 Hypophosphataemia 3 3 1 0 5 5 15 12 Palmar-plantar erythrodysaesthesia 9 0 4 1 57 13 45 II svndrome (PPES)

7.5.6 90-Day Safety Update

On January 23, 2017, Bayer submitted a 90-day safety update (as previously agreed to by the FDA). The new SAE reports from this update covered twenty five patients in Study 15982, twenty of whom were in the regorafenib arm. These reports did not reveal any new safety information.

8 Postmarket Experience

No new information pertinent to this application.

9 Appendices

9.1 Labeling Recommendations

The following table summarizes the significant recommended changes to the regorafenib label. As this review will be completed prior to the PDUFA goal date, some changes to the labeling may occur subsequent to the completion of this review that may

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

be addressed in an amendment to the clinical review. Rows in dark grey indicate the sections of the label with changes, rows in light grey indicate Bayer's proposal, and rows with no filling summarize FDA proposed changes to Bayer's proposal in bold type.

1 I ndications and Usage, 1.3 Addition of "Hepatocellular carcinoma (HCC) who have been previously treated witht:j I cl,, The indication in ref!ards to prior theraov is currently heinf! nef!otiated at this time. 2.2 Recommended doses and schedules Addition of' ~ (b)(4)1" FDA concmTed. Addition of 't (b)(4} I l" FDA proposed removing this sentence as insufficient info1mation was provided to support the safety of this recommendation. 5 Warnings and Precautions Bayer pooled data from the four placebo-controlled trials of regorafenib [CORRECT for mCRC, GRID for GIST, CONCUR for mCRC, and RESORCE (Study 15982)1 FDA concmTedl (b)(4)j I I "4800 patients treated with STIVARGA across all clinical trials;" FDA has asked for justification regarding the specific number of patients in all four trials. 6 Adverse Reactions "reactions (~20%) in patients receiving STIVARGA are pain (including gastrointestinal and abdoininal pain), HFSR, asthenia/fatigue,~ <4l diruThea, decreased appetite/food intake, 4 4 hypertension, infectionJ (bH >J dysphonia, (bH > hyperbilirnbinemia, fever, 4 4 mucositis, I (bH ~ weight loss, I (bH >j rash, ~ and nausea.,, FDA proposed "Pain, Asthenia/fatigue, PPES, Dianhea, Decreased appetite, HTN, Infection and infestations, Dysphonia, Mucositis, Pyrexia, Weight decrease, Constipation, Rash, Nausea, Hyperbilirnbinemia'' This change reflects FDA 's calculations ofadverse events; however, FDA will wait to receive Bayer's response (i.e., ifFDA agrees to the Bayer's methodology regarding the calculation of adverse events). 6.1 Clinical Trials Experience Bayer updated some of their adverse event incidence numbers for Colorectal Cancer and GIST. Additionally, a Hepatocellulru· Cru·cinoma subsection was added to the label. FDA agrees with the updated adverse event incidence numbers based on tables in the ADR JD with the exception of Colorectal cancer/Pain/Placebo/Grade~ 3. Regarding the Hepatocellulru· Carcinoma section, labeling review is ongoing. 8.5 Geriatric Use

FDA requested that Bayer analyze hype1tension for all grades and Grades > 3 for the pooled

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

data from the 4 randomized, placebo-controlled trials. 14 Clinical Studies ~ Hepatocellular Carcinoma (HCC) Addition of a new subsection Labeling review is ongoing.

9.2 Advisory Committee Meeting

No advisory committee meeting was held for this application.

9.3 Additional Tables

Table 30-Study 15982: Schedule of assessments (copied from submission)

Scrttning cycle 1 Cycle 2+ End or safety Follow-up (4Wffks) (4Wffks) trtoatm•nt follow-0p ~~~~~trne +4)> -UdaJ> 7 d•ys •. ~ !:~tt tniti•rlon ~durN Informed com.em. x lnctus.Jonfexduuin Criter?.l Cheeciedf x BonesconF MKllilili«ilrfaJl!lE<1f111ent X' S•fety •"Hsmema 12-le>d ECG• x x· x x Advexial>Es ~ Pnor and ~ant medications x Atclw!!g • x· x• x·

Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC

ScrHning C)'1:1e 2+ E.nd of S•l•ty Follow-up (4 weeks) trutrnont follow -<1p Visits could be c:cmbined as lon1J 35 +4<1¥ .!S:S@:SYntnt!. WM@ compl@~ w1!hio !he rune frames. Within Within Witnin Ooy1 Ooy1S. Ooy1 lSda.i> Ud3J' 7days •J4')o 3<1;¥ &3< g«t•1'c bi«nat1tm ' )(• )(• Pl•sm.a I<>< ~•c l>O>m:riOf> • )(' X' X' X' x• X' ECOG P"""""'""" Status )( x )( x PhysioaJ """"'"'°"°" x x x x Vlbl$VOS )( x )( x x x Blood P'•"u~ "'°"itomg x ~ x CBC wfth -l'l!floal' x x• x )( x• x Chemisuy & electrdyte panel' x x• x x x• x Urinalysis' x x• x x x• x Thyn:lid ~ff (TSl1.llM ~ 11M T4) x X' x x CoagtAoon pooe! (?r-INR. PTT. aPTT)•' x ~ ... x x Pregnancy i..st fd app6oalff.) x x~ X"' ~ GFR as~nmont x Alpha~•t"ll'OIM (AFP) x x x x Od>H Prom..oohnebo 5'1rnpiolQ • )( x x Pabint~ a..aliiyo'l.ile (EO-SOilnll FM:T-li: FACT-Hep =Functional Assg~nt of Cane"' Th•rapy-H•patobifi•ry: HCC = nepo:oc•ftul•r carcinoma; MRI = rnaQMtic resonal!* im>ging, PT·INR = pro1hrombin time -international normalized r.. tlo ; PTT = parial thromboplastm time; TNM = Tumor. node. metastasis; TSH = Thyroid-stimulobng hormone. a Hepatilis S su1face .mtigen .and antl-Hepa:ilks • 14 days. If tumo< usessments are availal>U! during the follow up periocl for subjects who discontinued study trea~t and have not ~91>nenoed PO. they shoulcl be"'"°"'"" in the CRF. c After 6 cycles, ECG can be perfonned based on In@ investiga,o(s discn!lion. .ECG is not required on Day 1 of Cycle I if done wilhin 7 days of starting srudy drug lte3.tml>nL d Adve

e Archival Biopsy for Biomarkers: An:::hiv.ll FFPE tumor tissue may b@supplied a:sa block or as pre-cut slid@-s. Pre-eutsidH should M ~shly prepar@d. unsta.1ned. and cut 1IO a !hiolI<>) and 10 mictOl'ls (n=5. If ~slble). Tumor tissue should be conected from subjects who dKI or did""' provide genetic consent Whole Blood for Siomarln~ -10 ml ol blood will be taken at eaell time point for plasmo. preparallon. For sul>jects who have pddi1ion. wttl5elinl!, thpble. Thfs Information win be recorded In ll>e CRF. h In al subj«,..,, 1he investigator and be.face any study related pmcl!dures .31'@ done-. so that 3ny intem.ction ~C! n th~ subjl!cts 3nd investigator or other he aJth care pmvid« will not influence !he tll'sponses to 1he questionnaires. Ouestlonn.aitff should be a.ry 3113lysis until end of study m tf req1nr-ed by nabon.J:I /instiMion.al ragUlation:s. pregnancy te.Sl sllould be pE-rformed for women of dlikl beanng potE'ntfal pnor to the study drug adm1nsstr.ltion at day one of "3ch cycle n Child-Pugh is not required at Cycle 1 Cay 1 if i• wa• ...ses.>ed wi1hin 7 days of starting swdy drug ~.;tmenl and if l.lboratory ev.alu.alions (bilirubln. albumin. and PT proklnqed or INR) we.re not compl•ted at Cycle 1 Day 1.

~~Table 31- Study 15982: AEs by PT (sorted by frequency, then alphabetical order)~~

~~no. of no. of PL o/oof o/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N=374~~

~~Palmar-plantar eiythrodysaesthesia 13 7 192 51~~

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

syndrome Diarrhoea 29 15 153 41 H ypettension 12 6 115 31 Decreased appetite 27 14 114 30 Fatigue 47 24 106 28 Aspartate aminotransferase 38 20 92 25 increased Blood bilirubin increased 31 16 91 24 Abdominal pain 30 16 79 21 Pyrexi a 13 7 74 20 Dysphonia 3 2 66 18 Constipation 21 11 65 17 Nausea 26 13 64 17 Ascites 31 16 58 16 Asthenia 18 9 56 15 Oedema peripheral 26 13 56 15 Alanine aminotransferase increased 21 11 54 14 H ypoalbuminaemia 14 7 52 14 Anaemia 21 11 51 14 Weight decreased 8 4 50 13 Abdominal pain upper 17 9 47 13 Vomiting 13 7 47 13 Back pain 17 9 45 12 General physical health deterioration 27 14 44 12 Cough 13 7 41 11 Muscle spasms 4 2 38 10 Hypophosphataemia 4 2 36 10 Platelet count decreased 2 1 34 9 Proteinuria 2 1 32 9 Stomati tis 4 2 31 8 Dyspnoea 15 8 28 7 Lipase increased 6 3 27 7 Alopecia 5 3 26 7 Hypokalaemia 5 3 26 7 Pain in extremity 6 3 26 7 Hypothyroidism 0 0 24 6 Insomnia 8 4 24 6 Headache 12 6 23 6 Blood alkaline phosphatase 8 4 22 6 increased

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Ganuna-glutamyltransferase 12 6 22 6 increased Malaise 5 3 22 6 Dry mouth 9 5 21 6 Hyponatraemia 6 3 21 6 Rash 14 7 20 5 Pmritus 14 7 19 5 Abdominal distension 10 5 18 5 Musculoskeletal pain 11 6 17 5 White blood cell count decreased 2 1 17 5 Chest pain 4 2 16 4 Epistaxis 2 1 16 4 Plemal effusion 11 6 15 4 Arthralgia 11 6 14 4 Bronchitis 1 1 14 4 Dyspepsia 4 2 14 4 H yperbilirnbinaemia 3 2 14 4 Blood thyroid stimulating ho1mone 3 2 13 3 increased Dizziness 8 4 13 3 Nasopharyngitis 2 1 13 3 Urinaiy tract infection 3 2 13 3 Anxiety 4 2 12 3 Blood lactate dehydrogenase 4 2 12 3 increased Hepatic encephalopathy 7 4 12 3 H yperkalaemia 7 4 12 3 H ypomagnesaemia 0 0 12 3 Myalgia 2 1 12 3 Amylase increased 0 0 11 3 Dysgeusia 1 1 11 3 Influenza like illness 6 3 11 3 Mucosa! inflammation 0 0 11 3 Neutrophil count decreased 2 1 11 3 Haemon-hoids 0 0 10 3 Orophaiyngeal pain 1 1 10 3 Erythema 1 1 9 2 Fall 4 2 9 2 Gastritis 3 2 9 2 Haematuria 2 1 9 2

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Hepatic failure 9 5 9 2 Hyperkeratosis I I 9 2 H ypocalcaemia 0 0 9 2 Hypoglycaemia 0 0 9 2 Pneumonia 2 I 9 2 Tinnitus 2 I 9 2 Urine bilirubin increased I I 9 2 Abdominal pain lower 0 0 8 2 Blood creatinine increased 5 3 8 2 Dry skin 5 3 8 2 Hyperglycaemia 5 3 8 2 Jaundice 3 2 8 2 Lymphocyte count decreased 2 I 8 2 Musculoskeletal chest pain 2 I 8 2 Thrombocytopenia 4 2 8 2 Upper respiratory tract infection 2 I 8 2 Bilirubin conjugated increased I I 7 2 Chills 0 0 7 2 Dehydration 0 0 7 2 Depression 4 2 7 2 Haemoglobin decreased 0 0 7 2 Haemoptysis 3 2 7 2 Hypotension 4 2 7 2 Oral candidiasis 0 0 7 2 Twnourpain 3 2 7 2 Bone pain 4 2 6 2 Dennatitis acneiform 4 2 6 2 Influenza I I 6 2 Leukocytosis I I 6 2 Muscular weakness 2 I 6 2 Pancreatitis 0 0 6 2 Paraesthesia 4 2 6 2 Pollakiw·ia 2 I 6 2 Atrial fibrillation 0 0 5 I Blood albumin decreased I I 5 I Cheilitis 0 0 5 I C-reactive protein increased I I 5 I Dysphagia 2 l 5 I Gastrointestinal pain I l 5 I Glossodynia 0 0 5 I

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Hepatic function abno1mal 5 3 5 I Mouth ulceration 2 I 5 I Peripheral senso1y neuropathy I I 5 I Skin lesion 2 I 5 I Tremor 0 0 5 I U1iicaria 0 0 5 I Ve1iigo 3 2 5 I Weight increased 0 0 5 I White blood cell count increased 0 0 5 I Wound 0 0 5 I Activated pa1iial thromboplastin I l 4 I time prolonged Blood urea increased I l 4 I Dysuria I I 4 l Encephalopathy 6 3 4 I Flank pain 2 l 4 I Flatulence 2 I 4 I Gingival pain 0 0 4 I Hepatic pain 2 l 4 I H yperhidrosis 0 0 4 I Hyperuricaemia 2 I 4 l Lm1g infection 2 I 4 I Monocyte cotmt increased 0 0 4 I Oedema 0 0 4 I Oesophageal varices haemo1Thage I I 4 I Pharyngitis l l 4 I Pneumonitis 0 0 4 I P01ial vein thrombosis 2 I 4 l Rash pustular I I 4 I Rectal haemon-hage 0 0 4 I Respirato1y tract infection 0 0 4 I Skin exfoliation 0 0 4 I Skin ulcer 0 0 4 I Tachycardia 0 0 4 I Tooth infection 0 0 4 l Upper gastrointestinal haemon-hage 4 2 4 I Varices oesophageal 0 0 4 I Abdominal infection 2 I 3 I Acute coronary syndrome 0 0 3 I Acute kidney injmy 3 2 3 I

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Alpha 1 foetoprotein increased 1 1 3 1 Angina pectoris 0 0 3 1 Blood glucose. increased 0 0 3 1 Candida infection 0 0 3 1 Cellulitis 1 1 3 1 Chest discomfort 1 1 3 1 Chronic gastritis 0 0 3 1 Dyspnoea exertional 2 1 3 1 Ear discomfort 0 0 3 1 Eiythema multifonne 0 0 3 1 Gastrointestinal infection 0 0 3 1 Gastrooesophageal reflux disease 3 2 3 1 Gingival bleeding 0 0 3 1 Gynaecomastia 1 1 3 1 Haematoma 1 1 3 1 Haemon-hoidal haemon-hage 0 0 3 1 Hepatic cin-hosis 2 1 3 1 Hyperaesthesia 0 0 3 1 Hyperammonaemia 0 0 3 1 Hyperthyroidism 0 0 3 1 H ypertriglyceridaemia 2 1 3 1 Lymphopenia 1 1 3 1 Neck pain 2 1 3 1 Neutropenia 0 0 3 1 Neutrophil count increased 0 0 3 1 Nocturia 0 0 3 1 Oral fongal infection 0 0 3 1 Pain 3 2 3 1 Protlu·ombin time prolonged 2 1 3 1 Rales 1 1 3 1 Rash maculo-papular 1 1 3 1 Rectal tenesmus 0 0 3 1 Scrotal e1ythema 0 0 3 1 Sepsis 0 0 3 1 Shock haemo1Thagic 0 0 3 1 Sleep disorder 2 1 3 1 Somnolence 4 2 3 1 Spinal pain 1 1 3 1 White blood cells urine positive 1 1 3 1 Abdominal discomfo1t 4 2 2 1

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Anal erosion 0 0 2 I Aphonia I I 2 I Asterixis 0 0 2 I Atrial flutter 0 0 2 I Autoinumme thyroiditis 0 0 2 I Bacterial test positive 0 0 2 I Blood potassium decreased 0 0 2 I Blood uric acid decreased 0 0 2 I Cholangitis I I 2 I Clu-onic kidney disease 0 0 2 I Contusion 3 2 2 I Crepitations 0 0 2 I Cystitis 2 I 2 I Decubitus ulcer 0 0 2 I Dental caries 0 0 2 I Depressed mood I I 2 I Dennal cyst 0 0 2 I Dennati tis 0 0 2 I Diabetes mellitus 2 I 2 I Diverticulum intestinal 0 0 2 I Duodenal ulcer 0 0 2 I Eczema 0 0 2 I Electrocardiogram QT prolonged 0 0 2 I Electrocardiogram T wave inversion 0 0 2 I Flushing 0 0 2 I Gait disturbance I I 2 I Gastritis erosive I I 2 I Gastroenteritis 0 0 2 I Gastrointestinal haemorrhage 4 2 2 I General physical condition abnormal 0 0 2 I Genital rash I l 2 I Glossitis 0 0 2 I Haematemesis I I 2 I Haematochezia 0 0 2 I Hepatic haemotThage 2 I 2 I Hepatomegaly l I 2 I Hepa.torenal syndrome 2 I 2 I Hiccups 2 l 2 I Hypoacusis 0 0 2 I Hypophagia I I 2 I

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Hypoproteinaemia 1 1 2 1 Iron deficiency 2 1 2 1 Jaundice cholestatic 1 1 2 1 Lethargy 2 1 2 1 Liver abscess 2 1 2 1 Malnutrition 1 1 2 1 Memory impairment 0 0 2 1 Nasal dryness 0 0 2 1 Nitrite urine present 0 0 2 1 Oral pain 0 0 2 1 Osteoporosis 0 0 2 1 Palpitations 2 1 2 1 Penile erythema 0 0 2 1 Peritonitis bacterial 1 1 2 1 Platelet count increased 0 0 2 1 Proctalgia 1 1 2 1 Productive cough 2 1 2 1 Rash erythematous 0 0 2 1 Rash papular 0 0 2 1 Red blood cell cow1t decreased 1 1 2 1 Renal failure 5 3 2 1 Rhinitis 0 0 2 1 Seizure 0 0 2 1 Sinus tachycardia 2 1 2 1 Spinal osteoarthritis 0 0 2 1 Subcutaneous abscess 0 0 2 1 Syncope 0 0 2 1 Thirst 0 0 2 1 Thyroxine free increased 0 0 2 1 Tinea pedis 1 1 2 1 Toothache 1 l 2 1 Tri-iodothyronine free increased 0 0 2 1 Urine analysis abnormal 0 0 2 1 Urine ketone body present 0 0 2 1 Urobilinogen urine increased 0 0 2 1 Vision blm1·ed 2 l 2 1 Acne 0 0 1 0 Acute hepatic failure 0 0 l 0 Acute sinusitis 0 0 l 0 Adenocarcinoma gastric 0 0 l 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no.of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Ammonia increased 0 0 1 0 Amnesia 0 0 1 0 Anal fissure 0 0 1 0 Anal fistula 0 0 1 0 Anal haemorrhage 0 0 1 0 Angular cheilitis 0 0 1 0 Anorectal discomfort 0 0 1 0 Anorectal infection 0 0 1 0 Aphasia 2 1 1 0 Aphthous ulcer 0 0 1 0 Aptyalism 0 0 1 0 Arrhythmia 1 1 1 0 Astigmatism 0 0 1 0 Atrioventi·icular block first degree 0 0 1 0 Bacteriuria 0 0 1 0 Benign neoplasm of thyroid gland 0 0 1 0 Benign prostatic hype1plasia 1 1 1 0 Bile duct stone 0 0 1 0 Blepharitis 0 0 1 0 Blister 1 1 1 0 Blood chloride increased 0 0 1 0 Blood phosphoms decreased 0 0 1 0 Blood pressure decreased 0 0 1 0 Blood pressure diastolic increased 0 0 1 0 Blood pressure increased 0 0 1 0 Blood sodium decreased 0 0 1 0 Blood thyroid stimulating ho1mone 0 0 1 0 decreased Bone marrow failw·e 0 0 1 0 Bradycardia 2 1 1 0 Brain neoplasm 0 0 1 0 Brain oedema 0 0 1 0 Breast mass 1 1 1 0 Breath sounds abnonnal 0 0 1 0 Bronchial obstmction 0 0 1 0 Calculus w·inary 0 0 1 0 Cancer pain 1 1 1 0 Cardiac faihu·e 0 0 1 0 Cataract operation 0 0 1 0 Cerebrovascular accident 0 0 1 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no.of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Cholecystitis 0 0 1 0 Circadian rhythm sleep disorder 0 0 1 0 Clostridium difficile colitis 0 0 1 0 Clostridium difficile infection 0 0 1 0 Coccydynia 1 1 1 0 Cold sweat 0 0 1 0 Colitis 0 0 1 0 Concussion 0 0 1 0 Conduction disorder 0 0 1 0 Confusional state 1 1 1 0 Conjunctiva! haemon-hage 0 0 1 0 Craniocerebral injwy 0 0 1 0 Cystitis noninfective 0 0 1 0 Death 0 0 1 0 Depressive symptom 0 0 1 0 De1111atitis allergic 0 0 1 0 Dennatitis atopic 0 0 1 0 Dennatitis bullous 0 0 1 0 Den11atitis exfoliative 0 0 1 0 De1111atitis psoriasifonn 0 0 1 0 Device related infection 0 0 1 0 Dianhoea haemorrhagic 0 0 1 0 Dislocation of vert.ebra 0 0 1 0 Disorientation 1 1 1 0 Diverticulum 0 0 1 0 Dmg eruption 0 0 1 0 Dty eye 0 0 1 0 Diythroat 0 0 1 0 Duodenal perforation 0 0 1 0 Duodenitis 0 0 1 0 Dysaesthesia 0 0 1 0 Eczema asteatotic 0 0 1 0 Electrocardiogram abno1111al 0 0 1 0 Embolism 0 0 1 0 Emotional disorder 0 0 1 0 Epidetnic pleurodynia 0 0 1 0 Epigastric disco1nfo1t 1 1 1 0 Epilepsy 0 0 1 0 Erectile dysfunction 1 1 1 0 Escherichia sepsis 0 0 1 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Eyelid oedema 0 0 1 0 Face oedema 0 0 1 0 Facial wasting 0 0 1 0 Febrile infection 0 0 1 0 Feeling abnormal 0 0 1 0 Feeling cold 1 1 1 0 Femur fracture 1 1 1 0 Fine motor skill dysfunction 0 0 1 0 Folliculitis 0 0 1 0 Food poisoning 0 0 1 0 Fungal skin infection 1 1 1 0 Funmcle 0 0 1 0 Gallbladder obstmction 0 0 1 0 Gallbladder pain 0 0 1 0 Gastritis haemonhagic 0 0 1 0 Genital candidiasis 0 0 1 0 Genital he1pes 0 0 1 0 Gingival oedema 0 0 1 0 Gingival ulceration 0 0 1 0 Gingivitis 0 0 1 0 Gout 2 1 1 0 Gravitational oedema 0 0 1 0 Groin pain 2 1 1 0 Haemobilia 0 0 1 0 Haemoglobin increased 0 0 1 0 Haemoffhage intracranial 0 0 1 0 Hepatic ischaeinia 0 0 1 0 Hepatitis acute 0 0 1 0 Hepatobilia1y disease 0 0 1 0 Hemiapain 0 0 1 0 He1pes zoster 0 0 1 0 Hot flush 3 2 1 0 H ypercalcaeinia 3 2 1 0 H ypenuetropia 0 0 1 0 Hypersensitivity 0 0 1 0 Hypersomnia 0 0 1 0 Hypertensive crisis 0 0 1 0 Hypertensive herut disease 0 0 1 0 Hyperventilation 0 0 1 0 H ypoaesthesia 1 1 1 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no.of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Hypogeusia 0 0 1 0 H ypoventilation 0 0 1 0 Hypovolaemic shock 0 0 1 0 Incontinence 0 0 1 0 Increased bronchial secretion 0 0 1 0 Infected neoplasm 0 0 1 0 Infection 4 2 1 0 Inflammation 0 0 1 0 Inguinal hernia 1 1 1 0 Intemational nonnalised ratio 1 1 1 0 increased Interstitial lung disease 0 0 1 0 Interve1tebral disc protmsion 0 0 1 0 In·itable bowel syndrome 0 0 1 0 Joint swelling 1 1 1 0 Jugular vein thrombosis 0 0 1 0 Lacrimation increased 0 0 1 0 Large intestine benign neoplasm 0 0 1 0 Laryngeal pain 0 0 1 0 Leukopenia 1 1 1 0 Ligament sprain 0 0 1 0 Lip pain 0 0 1 0 Lower respirato1y tract infection 0 0 1 0 Lung abscess 0 0 1 0 Lung hypoinflation 0 0 1 0 Melaena 1 1 1 0 MeningoIThagia 0 0 1 0 MenoIThagia 0 0 1 0 Migraine 0 0 1 0 Mouth haemoIThage 0 0 1 0 Muscle fatigue 0 0 1 0 Myasthenia gravis 0 0 1 0 Myocardial infarction 0 0 1 0 Nail disorder 0 0 1 0 Nasal congestion 0 0 1 0 Nasal he1pes 0 0 1 0 Nephrolithiasis 0 0 1 0 Nephrotic syndrome 0 0 1 0 Nervous system disorder 0 0 1 0 Neuralgia 0 0 1 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Neuropathy peripheral 1 1 1 0 Night sweats 1 1 1 0 Nipple pain 1 1 1 0 Non-cardiac chest pain 1 1 1 0 Ocular icterus 0 0 1 0 Odynophagia 1 1 1 0 Oesophageal candidiasis 1 1 1 0 Oesophagitis 0 0 1 0 Oliguria 0 0 1 0 Oral discomfort 0 0 1 0 Otitis extema 1 1 1 0 Otitis media 0 0 1 0 Pain of skin 0 0 1 0 Painful defaecation 0 0 1 0 Pallor 0 0 1 0 Pancreatitis acute 0 0 1 0 Pancytopenia 0 0 1 0 Paronychia 0 0 1 0 Parosmia 0 0 1 0 Pathological fractme 1 1 1 0 Pelvic fluid collection 1 1 1 0 Pelvic fracture 0 0 1 0 Pelvic pain 3 2 1 0 Peria1thritis 0 0 1 0 Periodontal disease 0 0 1 0 Periodontitis 0 0 1 0 Peripheral aite1y stenosis 1 1 1 0 Peripheral coldness 0 0 1 0 Peripheral motor neuropathy 0 0 1 0 Peripheral swelling 2 1 1 0 Periphlebitis 0 0 1 0 Peritonitis 2 1 1 0 Pharyngeal inflammation 0 0 1 0 Photopsia 0 0 1 0 Plantar erythema 0 0 1 0 Pneumonia aspiration 0 0 1 0 Polydipsia 0 0 1 0 Poor quality sleep 1 l l 0 Po1tal hypertensive gastropathy 0 0 l 0 Post-traumatic pain 0 0 l 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Presbyopia 0 0 1 0 Procedural pain 0 0 1 0 Protein total increased 1 1 1 0 Protein urine present 1 1 1 0 Prothrombin time sho1tened 0 0 1 0 P111ritus generalised 1 1 1 0 Pubic pain 0 0 1 0 Pulmonary congestion 0 0 1 0 Pulmonaiy embolism 1 1 1 0 Pulmonary oedema 0 0 1 0 Pulmonaiy venous thrombosis 0 0 1 0 Punctate keratitis 0 0 1 0 Plll'pura 1 1 1 0 Purulent discharge 0 0 1 0 Quadripai·esis 0 0 1 0 Rash generalised 0 0 1 0 Red blood cells urine 0 0 1 0 Red blood cells urine positive 1 1 1 0 Reflux gastritis 0 0 1 0 Renal impairment 0 0 1 0 Renal pain 0 0 1 0 Respiratory distress 0 0 1 0 Respirato1y failure 3 2 1 0 Respiratory tract infection viral 0 0 1 0 Restlessness 0 0 1 0 Retinal artery occlusion 0 0 1 0 Rhinon-hoea 4 2 1 0 Rib fracture 1 1 1 0 Scab 0 0 1 0 Scrotal swelling 0 0 1 0 Scrotal ulcer 0 0 1 0 Sebaceous hyperplasia 0 0 1 0 Septic shock 0 0 1 0 Skin fissures 0 0 1 0 Skin hype1pigmentation 0 0 1 0 Skin infection 0 0 1 0 Skin irritation 0 0 1 0 Skin reaction 1 1 1 0 Skin swelling 0 0 1 0 Spider naevus 0 0 1 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Spinal column injury 0 0 1 0 Spinal compression fracture 1 1 1 0 Squamous c.ell carcinoma of skin 0 0 1 0 Staphylococcal bacteraemia 0 0 1 0 Staphylococcus test positive 0 0 1 0 Status epilepticus 0 0 1 0 Stoma site haemorrhage 0 0 1 0 Streptococcal bacteraemia 0 0 1 0 Streptococcal infection 0 0 1 0 Subcutaneous haematoma 0 0 1 0 Supraventricular extrasystoles 1 1 1 0 Supraventricular tachycardia 0 0 1 0 Therapeutic embolisation 0 0 1 0 Thrombocytosis 0 0 1 0 Thrombosis 0 0 1 0 Thyroid neoplasm 0 0 1 0 Thyroxine increased 0 0 1 0 Tineacnu-is 0 0 1 0 Tongue discolow·ation 0 0 1 0 Tooth abscess 0 0 1 0 Tooth loss 0 0 1 0 Tracheal disorder 0 0 1 0 Tracheitis 0 0 1 0 Traumatic haemotThage 0 0 1 0 Tri-iodothyronine free decreased 0 0 1 0 Tumour associated fever 0 0 1 0 Umbilical hemia 0 0 1 0 Upper extre1nity mass 0 0 1 0 Ureterolithiasis 0 0 1 0 U11na1y incontinence 1 1 1 0 Urine output decreased 0 0 1 0 Urine transitional cells present 0 0 1 0 Urosepsis 0 0 1 0 Vasculitis 1 1 1 0 Venous thrombosis limb 0 0 1 0 Viral infection 0 0 1 0 Visual field defect 0 0 1 0 Wolff-Parkinson-White syndrome 0 0 1 0 Wound infection 0 0 1 0 Accident 1 1 0 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Ageusia 1 1 0 0 Arthritis 2 1 0 0 Atelectasis 1 1 0 0 Bile duct stenosis 4 2 0 0 Blood creatine phosphokinase 1 1 0 0 increased Blood triglycerides increased 1 1 0 0 Blood uric acid increased 1 1 0 0 Breast pain 1 1 0 0 Bw11ing sensation 1 1 0 0 Cardiac arrest 1 1 0 0 Cerebral haematoma 1 1 0 0 Change of bowel habit 1 l 0 0 Cholestasis 2 l 0 0 Conjunctivitis 2 l 0 0 Deep vein thrombosis 2 l 0 0 Dyskinesia 1 1 0 0 Eating disorder 1 1 0 0 Eye haemotThage l l 0 0 Eye pain 1 l 0 0 Fibrin D dimer increased 1 l 0 0 Gastric disorder 1 l 0 0 Gastric haemoffhage 1 l 0 0 Genital cyst 1 1 0 0 Glycosuria 1 1 0 0 Haemaithrosis l l 0 0 Haemoglobinuria l l 0 0 Haemoll'hagic ai1aemia 1 l 0 0 Hemiparesis 1 l 0 0 Hepatitis B 1 l 0 0 Hepatocellular injwy 1 1 0 0 Hordeolum 1 1 0 0 Hypematraemia l l 0 0 Hyperthermia l l 0 0 Ileus 1 l 0 0 Infusion related reaction 1 l 0 0 Ingrown hair 1 l 0 0 Intennittent claudication 1 l 0 0 Intestinal obstmction 1 1 0 0 Intra-abdominal haemoffhage 2 1 0 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

Iron deficiency anaemia 1 1 0 0 Limb discomfort 1 1 0 0 Lower gastrointestinal haemon-hage 1 1 0 0 Lymph node pain 2 1 0 0 L ymphadenitis 1 1 0 0 Metastases to lung 1 1 0 0 Multiple organ dysfunction 1 1 0 0 syndrome Musculoskeletal discomfort 1 1 0 0 Musculoskeletal stiffness 3 2 0 0 Nodule 1 1 0 0 Obstrnction gastric 1 1 0 0 Oesophageal carcinoma 1 l 0 0 Oesophageal haemon-hage 1 1 0 0 Oesophageal stenosis 1 1 0 0 Oesophageal ulcer 1 1 0 0 Onychomalacia 1 1 0 0 Oropharyngeal candidiasis 1 1 0 0 Orthostatic hypotension 1 1 0 0 Overdose 1 l 0 0 Palmar erythema 1 1 0 0 Papule 1 1 0 0 Periorbital haematoma 1 1 0 0 Petechiae 2 1 0 0 Plantar fasciitis 1 1 0 0 Pleural infection bacterial 1 1 0 0 Pubis fracture l l 0 0 Pulmonaiy thrombosis 1 1 0 0 Renal aiie1y arteriosclerosis 1 1 0 0 Sciatica 2 1 0 0 Sinus congestion 1 1 0 0 Skin disorder 1 1 0 0 Tendonitis 1 1 0 0 Thennal bum l l 0 0 Twnour haem01Thage 1 1 0 0 Urinary retention 1 1 0 0 Uterine leiomyoma 1 1 0 0 Vena cava embolism 1 1 0 0 Visual acuity reduced 1 1 0 0 Vitamin D deficiency 1 1 0 0

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

~~no. of no. of PL %of O/oofPL Regorafenib PT Grades 1-5 Regorafenib N=193 Grades 1-5 N = 374~~

~~Vocal cord paralysis 1 1 0 0 Vulvitis 1 1 0 0 Waist circumference increased 1 1 0 0 Wound infection bacterial 1 1 0 0~~

~~REFERENCES~~

1. Siegel RL, Miller KO, Jemal A. Cancer statistics, 2016. CA Cancer J Cl in 2016;66:7-30. 2. El-Serag HB. Hepatocellular carcinoma. N Engl J Med 2011 ;365:1118-27. 3. Bruix J, Qin S, Merle P, et al. Regorafenib for patients with hepatocellular carcinoma who progressed on sorafenib treatment (RESORCE): a randomised, double blind, placebo-controlled, phase 3 trial. Lancet 2017;389:56-66. 4. Llovet JM, Decaens T, Raoul JL, et al. Brivanib in patients with advanced hepatocellular carcinoma who were intolerant to sorafenib or for whom sorafenib failed: results from the randomized phase Ill BRISK-PS study. J Clin Oncol 2013;31 :3509-16. 5. Zhu AX, Kudo M, Assenat E, et al. Effect of everolimus on survival in advanced hepatocellular carcinoma after failure of sorafenib: the EVOLV E-1 randomized clinical trial. JAMA 2014;312:57-67. 6. Zhu AX, Park JO, Ryoo BY, et al. Ramucirumab versus placebo as second-line treatment in patients with advanced hepatocellular carcinoma following first-line therapy with sorafenib (REACH): a randomised, double-blind, multicentre, phase 3 trial. Lancet Oneel 2015;16:859-70. 7. Wilhelm SM, Dumas J, Adnane L, et al. Regorafenib (BAY 73-4506): a new oral multikinase inhibitor of angiogenic, stromal and oncogenic receptor tyrosine kinases with potent preclinical antitumor activity. Int J Cancer 2011 ;129:245-55. 8. Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advanced hepatocellular carcinoma. N Engl J Med 2008;359:378-90. 9. Cheng AL, Kang YK, Chen Z, et al. Efficacy and safety of sorafenib in patients in the Asia-Pacific region with advanced hepatocellular carcinoma: a phase Il l randomised, double-blind, placebo-controlled trial. Lancet Oneel 2009;10:25-34. 10. Shah DR, Shah RR, Morganroth J. Tyrosine kinase inhibitors: their on-target toxicities as potential indicators of efficacy. Drug Saf 2013;36:413-26. 11. Lupo G, Caporarello N, Olivieri M, et al. Anti-angiogenic Therapy in Cancer: Downsides and New Pivots for Precision Medicine. Front Pharmacol 2016;7:519. 12. Fong ZV, Tanabe KK. The clinical management of hepatocellular carcinoma in the United States, Europe, and Asia: a comprehensive and evidence-based comparison and review. Cancer 2014;120:2824-38.

~~Reference ID: 4081175 Clinical Review Lorraine Pelosof sNDA 203085 Stivarga®/regorafenib in 2nd line HCC~~

13. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127:2893- 917. 14. Balogh J, Victor D, 3rd, Asham EH, et al. Hepatocellular carcinoma: a review. J Hepatocell Carcinoma 2016;3:41-53. 15. Bruix J, Sherman M, American Association for the Study of Liver D. Management of hepatocellular carcinoma: an update. Hepatology 2011;53:1020-2. 16. Llovet JM, Bru C, Bruix J. Prognosis of hepatocellular carcinoma: the BCLC staging classification. Semin Liver Dis 1999;19:329-38. 17. Toyoda H, Kumada T, Kiriyama S, et al. Comparison of the usefulness of three staging systems for hepatocellular carcinoma (CLIP, BCLC, and JIS) in Japan. Am J Gastroenterol 2005;100:1764-71. 18. Mazzaferro V, Regalia E, Doci R, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693- 9. 19. Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepatocellular carcinoma. Semin Liver Dis 2010;30:52-60. 20. Patchett N, Furlan A, Marsh JW. Decrease in tumor enhancement on contrast- enhanced CT is associated with improved survival in patients with hepatocellular carcinoma treated with Sorafenib. Jpn J Clin Oncol 2016;46:839-44.

~~83 Reference ID: 4081175 Clinical Investigator Financial Disclosure Review Template~~

~~Application Number: sNDA 203085 Submission Date(s): October 6, 2016; October 30, 2016 Applicant: Bayer HealthCare Pharmaceuticals Inc. Product: regorafenib~~

Reviewer: Lorraine Pelosof Date of Review: April 6, 2017 Covered Clinical Study (Name and/or Number): Study 15982 “A randomized, double blind, placebo-controlled, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib”

Was a list of clinical investigators provided: Yes No (Request list from applicant) Total number of investigators identified: 971 Number of investigators who are sponsor employees (including both full-time and part-time employees): 0

Number of investigators with disclosable financial interests/arrangements (Form FDA 3455): 4 If there are investigators with disclosable financial interests/arrangements, identify the number of investigators with interests/arrangements in each category (as defined in 21 CFR 54.2(a), (b), (c) and (f)): Compensation to the investigator for conducting the study where the value could be influenced by the outcome of the study: 0 Significant payments of other sorts: 4 Proprietary interest in the product tested held by investigator: 0 Significant equity interest held by investigator in sponsor of covered study: 0 Is an attachment provided with details Yes No (Request details from of the disclosable financial applicant) interests/arrangements: Is a description of the steps taken to Yes No (Request information minimize potential bias provided: from applicant) Number of investigators with certification of due diligence (Form FDA 3454, box 3) 971 Is an attachment provided with the Yes No (Request explanation reason: from applicant)

~~Reference ID: 4081175~~

Discuss whether the applicant has adequately disclosed financial interests/arrangements with clinical investigators as recommended in the guidance for industry Financial Disclosure by Clinical Investigators.1 Also discuss whether these interests/arrangements, investigators who are sponsor employees, or lack of disclosure despite due diligence raise questions about the integrity of the data: - If not, why not (e.g., study design (randomized, blinded, objective endpoints), clinical investigator provided minimal contribution to study data) - If yes, what steps were taken to address the financial interests/arrangements (e.g., statistical analysis excluding data from clinical investigators with such interests/arrangements) Briefly summarize whether the disclosed financial interests/arrangements, the inclusion of investigators who are sponsor employees, or lack of disclosure despite due diligence affect the approvability of the application.

~~Please refer to Section 3 of the Clinical Review.~~

~~1 See [web address].~~

Reference ID: 4081175 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------LORRAINE C PELOSOF 04/06/2017

~~STEVEN J LEMERY 04/07/2017~~

~~Reference ID: 4081175 CENTER FOR DRUG EVALUATION AND RESEARCH~~

~~APPLICATION NUMBER:~~

~~203085Orig1s007~~

~~CHEMISTRY REVIEW(S)~~

~~ !"#$ %& '()01 2"# # 2# ) ~~

~~34 5%6# 7881 549@ABACDEFAAG~~

H4 7I1 PQR '(S )6'7#18 UV% 7I1 T 8 7I1 # % # '(# # W## ArsXtis@Aurv XY`a`bcdFeffdghgbi pq BAsXtis@Aur AGswxys@Aur @@sgs@AuG @Dscbs@AuG BusXtis@Aurvv vpcYiux@xdd`baFeh` `xbcigvvpcYi@x@xdd`baFeh` `xbcig

4 P#QVV1 )9 qbg`b`tci`xbxYFi`ycYacYgaxYcgb`cdgiAhaxYigiYgcihgbixfci`gbi `igfcixtgddedcY tcYt`bxhcxcygggbfYgy`xe ddiYgcig`i (b) (4) PIQ% #)0# PQ8 07881 9bxbg f4 '(gSu h4 ) #'( Si`y` `xbxXbtxdxadpYxeti @iXp@ j4 5#1# %%88# S kcdgYgcdicYgpcYhctgei`tcd lbte uAAkcdgYkdye peXekxmnuD o`ffcbdwAGnCusAnuDu

~~p4 S 'q 78# 5#1 #V1 7 0 '%1 # T) Fi`ycYacr cdgi`dh txcig Aha xYcd m wx~~

s4 )01#5#1# 77I# S tFqwSYgaxYcgb` gh`tcdbchgSsusvustdxYxsBsiY`dexYxhgiddfgbddwtcYchxddx ch`bxsBsdexYxfgbxmdwswshgiddfdY``bgs@stcYxmch`ghxbxdYcig

~~yxdgtedcYxYhedcS@uuDdwXBz@X yoSDAAeCB~~

{4 & # S Fi`ycYac` c|`bc g`b`ìxY`b`tcigxYigiYgcihgbixfci`gbi ìhgic ici`ttxdxYgticdtcbtgYx cygggbfYgy`xe ddiYgcigìdexYxfdY`h``bgsxmcd`fdci`bscb`Y`bxigtcbsc gtghxigYcfdcbcbi`s }~igYcfdcb`qF`dsidfgcbcbi`s~igYcfde

~~34788 '# 1 Sbxbg~~

~~334) Sbxbg~~

3H46q711# S g~`tctd wiq@ABACDEFAAG gg| ix`btdegcicxYYgaxYgbìcdg AhaxYigiYgcihgbixfci`gbi ì gfcixtgddedcYtcYt`bxhcxcygggbfYgy`xe ddiYgcigì (b) (4) gcic effxYi`bai` g`tctdc gY`ygYxh FieduDnC@cd x|bxbc ~FX~cYcbxh`gxedgsd`bfdctgxstxbiYxddg hedi`tgbigYadxcdpc gB iedxYgaxYcgb``bfci`gbi ìgfcixtgddedcYtcYt`bxhcxddx`ba xYcgb` iYgcihgbiepci`gbi xcygfYxaYg gxb xYcgb`xtehgbigYc`xdxa`tcdfYxaYg `xbcttxY`baixigYc`xdxad ÿ ÿÿÿÿÿÿ !ÿ"!!#$%ÿ&$' ÿ

()! !%ÿ!ÿ ! 0ÿ )ÿ!!#$ÿ!ÿ1'($ÿ23#ÿÿ4ÿ!#056 #ÿ3()57ÿ # 3ÿ!ÿ!#0560ÿ ÿ!(ÿ )!ÿ!!#$ÿ85ÿ#(ÿ0'9ÿ!ÿ@3ÿÿ!9ÿA@ÿ"ÿ09%ÿ(9('ÿ"77Bÿ@3ÿ#@ÿ%ÿ1'23ÿ$3 ÿ3211! ÿ (!ÿ"CD%ÿ!ÿ0# ('ÿ1'($ÿ $' 3ÿ1'23ÿCD7ÿE!ÿ3 209ÿFÿ )!ÿ3ÿ#ÿ()#ÿ ÿ )ÿ11!0ÿ01ÿ"85ÿ3ÿAÿ $' 3ÿÿAÿ5Bÿ!''9%7ÿ&)ÿ1'($ÿ $' ÿ3ÿ0# ('ÿ ÿ )ÿ11!0ÿ0!2Bÿ )23ÿ332!#ÿ$'#0#7ÿGHIPQÿSTUVWÿ XY`GaYb`cÿdefghÿGHIPQÿYipqrHBÿ#0ÿs'( !#(ÿt 3 3ÿ#0ÿD3ÿ 1! ÿE!53ÿ3ÿ1!00ÿ#ÿu02'ÿÿ3( #ÿ TvwvTvS7ÿ ÿ s((9ÿ3xtyÿÿ3ÿ2! )!ÿ3211! 0ÿ$9ÿ0 ÿ!5ÿ 209ÿAF8Bÿ#ÿ2#(# !''0ÿ1#ÿ'$'Bÿ52' (# !Bÿ )3ÿÿ3 9ÿ3 209ÿ#ÿ1 # 3ÿ )ÿ)1 (''2'!ÿ(!(#5ÿ"DD%7ÿE!ÿ3 209ÿAF8ÿ )!ÿ3ÿ#ÿ()#ÿ ÿ )ÿ 11!0ÿ01ÿ"85ÿ3ÿAÿ $' 3ÿÿAÿ5Bÿ!''9%ÿÿ#ÿ1'($ÿ"#ÿ$'#0#%7ÿGHIPQÿSTUÿPPifPI ÿGHIPQÿYipqrHÿ #0ÿs'( !#(ÿt 3 3ÿE!53ÿ!ÿ1!00ÿ#ÿu02'ÿÿ3( #ÿTvwvTvW7ÿ ÿ &)ÿ1!130ÿ210 0ÿ1(@ÿ#3! ÿ!ÿ!!#$ÿ $' 3ÿ3ÿ1!00ÿ#ÿ((!0#(ÿ )ÿ )ÿ1)93(#3ÿ'$'#ÿ!2'ÿ " %BÿÿDE ÿ787ÿ&)ÿÿ3ÿ1!00ÿ3ÿ$ )ÿÿ0! Bÿ('#ÿ#0ÿ## 0ÿ'Bÿ#0ÿ#ÿ !2( 2!0ÿ!02( ÿ $'#ÿ"%ÿ!5 7ÿ&)ÿÿ!5 ÿ''3ÿ )ÿ20#(ÿ3 ÿ! )ÿ#ÿ20#(ÿ!ÿ#023 !9Bÿ!0#ÿ 2' !9ÿ 2$533#3ÿ#ÿs'( !#(ÿE!5 ÿÿD# # ÿÿ$'#Bÿ3320ÿ$9ÿ )ÿD# !ÿ!ÿt!2ÿs'2 #ÿ#0ÿ 3!()ÿ#ÿy1!'ÿ 7ÿ&)ÿ210 0ÿÿ03ÿ# ÿ()#ÿDuDÿ!'# ÿ3( #3ÿ""W%B"w%ÿÿ"SS%%7ÿ( #3ÿSvSÿÿSvWÿ)ÿ$#ÿ(5$#0ÿ 2#0!ÿ3( #ÿSÿqÿGIppheiPGHqrgiÿgfPÿgfPhef7ÿxÿ()#3ÿ)ÿ$#ÿ50ÿ ÿ )ÿ(2!!# ÿ11!0ÿ ÿ#ÿ 3( #3ÿSvSÿÿSvW7ÿ&)ÿ210 0ÿÿ!ÿ((9ÿ3xtyÿ3ÿ((1 $'ÿ!ÿDuD7ÿ ÿ yÿ!23 ÿ!ÿ( !('ÿ('23#ÿ!5ÿ )ÿ#!#5# 'ÿ33335# ÿ"sy%ÿ!2!5# ÿ#ÿ((!0#(ÿ )ÿÿDE ÿ 7"$%ÿÿ )ÿEtyÿ20#(ÿ!ÿ#023 !9Bÿ0 0ÿ2'9ÿFFÿÿ# '0ÿds#!#5# 'ÿy33335# ÿÿ25#ÿt!2ÿ#0ÿ C'(3ÿy11'( #3Beÿ3ÿ1!00ÿ#ÿ3( #ÿSvSWvSÿÿ ''#ÿ2$533#ÿfÿ"FGghSTTc7ÿC9!ÿ!23 3ÿ ( !('ÿ('23#ÿ!5ÿsyÿ2#0!ÿ )ÿ$33ÿ ) ÿ )ÿs1( 0ÿ# !02( #ÿD#(# ! #ÿ"sD%ÿÿ )ÿ( ÿ5 9ÿ#ÿ )ÿ2 (ÿ#!#5# ÿ3ÿ$'ÿ )ÿ )!3)'0ÿÿ711$7ÿE2! )!5!Bÿ ÿ3ÿ# (1 0ÿ ) ÿ )3ÿ1!130ÿ( #ÿ2'0ÿ # ÿ3#(# '9ÿ( ÿ )ÿ2' 9ÿÿ)25#ÿ#!#5# 7ÿC9!3ÿ!23 ÿ!ÿÿ( !('ÿ('23#ÿ3ÿ((1 $'7ÿ ÿ C30ÿ#ÿ )ÿ3211! #ÿ0 ÿ1!00Bÿs((9ÿ3xtyÿ3ÿ!(55#00ÿ!ÿ11!'ÿ!5ÿÿDuDÿ 3 #01# 7ÿ ÿ SwvÿbqfihIjeqfjÿkÿYiiq ifPgHeqfj4ÿ &)3ÿ((9ÿ3211'5# ÿ3ÿ!(55#00ÿ!ÿ11!'ÿ!5ÿÿDuDÿ1!31( 7ÿ ÿ Svÿbq ifHjgileieifieijÿHqÿmiÿbqfniQiPÿHqÿppheigfH4ÿ##ÿ ÿ STvÿore grQÿYineiir4ÿÿ !#6ÿ ((BÿDuDÿ !BÿC!#()ÿBÿt3#ÿÿ3 u!@ #ÿy( 3ÿBÿp(ÿÿ(9('ÿt!2ÿ !02( 3ÿ"pt%Bÿp(ÿÿ)!5(2 ('ÿq2' 9ÿ"pq%ÿ ÿ

ÿ } ÿvuÿ ÿrstuvwsÿyzÿ{s||}ÿ~ÿ rstuvwsÿyzÿÿ|ÿszzÿsutvuvÿsÿ syÿsusuÿ zzzzzz {s||}ÿ~ ÿ|vrstuvwsÿyzÿ{s||}ÿ~ÿ }uÿzzÿÿ~ ÿ SvÿGiiqfPgrQÿYineiir4ÿÿ 53)ÿ )()!BÿC!#()ÿD)BÿC!#()ÿBÿt3#ÿÿ3 u!@ #ÿy( 3ÿBÿptBÿpqÿ ÿ ÿ } ÿvuÿ ÿ{}u ÿ{} }}| }tÿ~ÿ {}u ÿ ÿ|ÿszzÿsutvuvÿsÿ syÿsusuÿ zzzzzzÿ {} }}| }tÿ~|v{}u ÿ{} }}| }tÿ~ÿ }uÿzzÿÿ~

~~ÿ ÿ ÿÿÿÿÿÿ !ÿ"!!#$%ÿ&$' ÿ~~

()(ÿ1223224356ÿ 78ÿ@ABCDEFG5Hÿ75IFE4A6PF5ÿ &QÿRSSTÿUVWXÿÿQUÿ$#ÿY!Y!`ÿ#ÿSS!`#Sÿ Qÿ!a# Uÿ!SQ`ÿ QÿWbÿ!aÿ Qÿ&TYÿ cÿd #ÿ"e! #ÿ UY#UUÿb#'T%ÿ#ÿfgd!ghÿ#`ÿ Qÿ!gUVWXÿ&TYÿiÿa #ÿ#ÿgpq'ghrÿ ÿ b#ÿgpq'ghsÿtWXÿ!# `ÿuA26ÿvEABCÿw32PD5A6PF5ÿ!ÿ!!#$ÿ!ÿ Qÿx ! a# ÿÿY # Uÿ Qÿ ! ÿ(b) (4) QY S''q'!ÿS!S#aÿ"ycc%ÿQÿQÿ$#ÿY!qU'Tÿ ! `ÿ QÿU!#$sÿ#`ÿ#ÿfgpq#ghÿtWXÿ!# `ÿ EA5ÿwEGDÿw32PD5A6PF5ÿ!ÿ!!#$ÿ!ÿ Qÿ ! a# ÿÿQY S''q'!ÿS!S#arÿ ÿ iT!ÿUq$a `ÿ 3G326ÿIFEÿG4P22PF5ÿFIÿFE6PF52ÿFIÿA5ÿ1PBA6PF5ÿ6Fÿ7wÿÿ#ÿgYghÿÿEA56ÿ FP5Dÿ 3dP3eÿf3663Eÿ` `ÿfgbS ghrÿiT!ÿUq$a `ÿ ''#ÿq$aUU#ÿgÿ"hiwj%ÿ#ÿhgbS ghÿÿ ''#ÿq$aUU#ÿgÿ"hikwj%ÿ#ÿgbS ghrÿ ÿ 778ÿEFF23Hÿ(A5D32ÿ lÿXÿ#ÿ#`S #ÿ!ÿ !ÿ"!!#$%ÿ&$' sÿfaÿ!ÿ Qÿ ! a# ÿÿY # Uÿ QÿQY S''q'!ÿ S!S#aÿ"ycc%ÿQÿQÿ$#ÿY!qU'Tÿ ! `ÿ Qÿÿ ÿ !ÿÿ(b) (4) ÿ 7778ÿwA6AÿG4P663Hÿ6FÿGFE6ÿ63ÿEFF23Hÿ(A5D32ÿ ''#ÿq$aUU#ÿgÿ"gbS gh%ÿÿÿ#S'qÙÿÿQTY!'#nÿ ÿ Qÿcqaq' ÿ !oUÿpq`ÿ"j8%ÿ ÿ ''#ÿq$aUU#ÿgÿÿgrÿ&Qÿ&$'ÿÿc# # Uÿÿ Qÿcqaq' ÿ !oUÿpq`ÿ'U Uÿ QÿSqaq' ÿS# # Uÿÿ ''#ÿq$aUU#ÿgÿ"hgbS gh%ÿÿ ''#ÿq$aUU#ÿgÿ"gbS gh%sÿÿ#S'qÙÿQTY!'#nUÿ ÿ Qÿ''#ÿ US #Uÿÿ ''#ÿq$aUU#ÿgÿÿ ''#ÿq$aUU#ÿgqÿ lÿr8j8sÿd`q'ÿrrfqÿR#!#a# 'ÿX#'TUUÿ" ''#ÿq$aUU#ÿg%ÿ lÿr8j8jjÿd`q'ÿrfqÿt$'#ÿ lÿr8r8ÿd`q'ÿrrrqÿ q`Tÿ Y! Uÿÿc# !''`ÿc'#S'ÿ qÙÿ! ## ÿ ÿ Qÿc'a`ÿu#`S #vyccwÿ lÿr8r8xÿd`q'ÿrrrqÿ q`Tÿ Y! Uÿÿy#S# !''`ÿc'#S'ÿ qÙÿ ÿ &QÿUS #Uÿ'U `ÿ$ÿ`hq 'TÿUqYY! ÿ QÿYY!'ÿÿSSTÿUVWXÿÿ!aÿÿcdcÿY!UYS rÿt!ÿ $! Tÿ QÿUqYY! #ÿcdcÿ` ÿUÿ# ÿ!Y!`qS`ÿQ!ÿ$q ÿUÿ'$'ÿ#ÿ ''#ÿq$aUU#ÿgÿ"hgbS gh%ÿÿ ''#ÿq$aUU#ÿgÿ"gbS gh%rÿ ÿ d3EAÿzdAGA6PF5{ÿXSSY $'ÿ ÿ 7|8ÿ P2Cÿ122FBPA63HÿeP6ÿ63ÿEFF23Hÿ(A5D32ÿA5Hÿ74AB6ÿ6FÿEFHGB6ÿ}GAP6~ÿA5HÿA6P356ÿAI36~ÿ &QÿYY'S# ÿQUÿY!``ÿUqS# ÿ#`ÿ`hq ÿ` ÿ ÿS#S'q`ÿ Q!ÿ!ÿ#ÿcdcÿUUqUÿUUS `ÿ Qÿ !Saa#`#ÿYY!'ÿÿRSSTÿUVWXÿÿ!aÿÿcdcÿY!UYS rÿÿ ÿ

~~ÿ CENTER FOR DRUG EVALUATION AND RESEARCH~~

~~APPLICATION NUMBER:~~

~~203085Orig1s007~~

~~STATISTICAL REVIEW(S)~~

~~U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research Office of Translational Sciences Office of Biostatistics~~

~~S TATISTICAL R E V I E W A N D E VALUATION ADDENDUM~~

NDA/BLA #: 203085 Serial #: 007 Drug Name: Stivarga (Regorafenib) Indication: The treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4) Applicant: Bayer HealthCare Pharmaceuticals Received Date: October 6, 2016; October 31, 2016 PDUFA Date: April 30, 2017 Review Type: Priority Biometrics Division: Division of Biometrics V Statistical Reviewer: Xiaoping (Janet) Jiang, Ph.D. Concurring Reviewers: Lisa Rodriguez, Ph.D., Team Leader Kun He, Ph.D., Associate Director Medical Division: Division of Oncology Products 2 Clinical Team: Lorraine Pelosof, M.D., Ph.D., Clinical Reviewer Steven Lemery, M.D., Clinical Team Leader Patricia Keegan, M.D., Division Director Project Manager: Anuja Patel, M.P.H.

~~Keywords: Stratified log-rank test, Kaplan-Merier method, Cox regression~~

~~Reference ID: 4090305 Table of Contents TABLE 1 RESULT OF OS PRIMARY ANALYSIS...... 3 TABLE 2 RESULT OF OS SENSITIVITY ANALYSIS (EXCLUDING PATIENTS FROM ALL SITES IN CHINA)...... 4~~

~~Reference ID: 4090305~~

This addendum is to Dr. Xiaoping (Janet) Jiang’s statistical review for efficacy supplement of New Drug Application 203085S007 (dated on April 6, 2017). This addendum contains an exploratory sensitivity analysis of OS excluding sites in China.

On October 6, and October 31, 2016, the applicant submitted a sNDA to seek an approval of regorafenib for the proposed indication ‘treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4) ’. In the submissions, the applicant provided clinical data and results from Study 15982 (RESORCE) entitled ‘A randomized, double blind, placebo-controlled, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib’, and other studies.

In Study RESORCE, a total of 573 eligible patients were randomized in a 2:1 ratio to receive regorafenib plus best supportive care (BSC) or match placebo plus BSC. The randomization was stratified by geographical region (Asia versus rest of world), Eastern Cooperative Oncology Group (ECOG) performance status (0 versus 1), Alpha-fetoprotein level (<400 ng/mL versus ≥400 ng/mL), extrahepatic disease (presence versus absence), and macrovascular invasion (presence versus absence). The primary endpoint was overall survival (OS). The primary analysis was a stratified log-rank test. For details regarding the design, data analyses, and results of Study RESORCE, please refer to Dr. Xiaoping (Janet) Jiang’s statistical review (dated on April 6, 2017).

On April 12, 2017, the applicant submitted an amendment to the supplemental new drug application (NDA203085S007). In the submission, the applicant provided some additional findings that were not included in the clinical study report (CSR) for the pivotal clinical study RESORCE. The majority of the findings (data corrections or data additions) were from sites in China and Taiwan. On April 20, 2017, in the Teleconference with FDA, the applicant informed FDA that there was a small efficacy difference due to the data from one or two sites in China. There are 27 sites from China in RESORCE with a total of 137 patients. Based on the review team’s internal discussion, this statistical reviewer conducted an exploratory OS sensitivity analysis by excluding 137 patients from all 27 sites in China. Table 1 summarizes OS primary analysis that was presented in Dr. Xiaoping (Janet) Jiang’s April 6, 2017 statistical review.

Table 1 Result of OS primary Analysis Placebo + BSC Regorafenib + BSC n=194 n=379 Number of Event (%) 140 (72.2) 233 ( 61.5) Median OS in Months (95% CI) 7.8 (6.3, 8.8) 10.6 (9.1, 12.1) Hazard Ratioa (95%CI) 0.63 (0.50, 0.79) P-valuea <0.0001 a stratified by line of therapy (3rd-line vs. 4th-line or beyond), geographical region (Asia vs. rest of world (ROW)), ECOG-PS (0 vs. 1), Alpha-fetoprotein level (<400 ng/mL vs. ≥400 ng/mL), presence vs. absence of extrahepatic disease, and presence vs. absence of macrovascular invasion.

~~Reference ID: 4090305 Table 2 summarizes the exploratory OS sensitivity analysis.~~

Table 2 Result of OS Sensitivity Analysis (Excluding patients from all sites in China) Placebo + BSC Regorafenib + BSC n=145 n=291 Number of Events (%) 108 (37.1) 183 (62.9) Number of Censored (%) 37 (25.5) 108 (37.1) Median OS in months (95% CI) 8.3 (6.8, 9.3) 10.9 (9.1, 13.2) Hazard ratioa (95%CI) 0.62 (0.48, 0.80) a stratified by line of therapy (3rd-line vs. 4th-line or beyond), geographical region (Asia vs. rest of world (ROW)), ECOG-PS (0 vs. 1), Alpha-fetoprotein level (<400 ng/mL vs. ≥400 ng/mL), presence vs. absence of extrahepatic disease, and presence vs. absence of macrovascular invasion.

~~Reviewer’s Comment: 1. As shown in Table 2, the results of the OS sensitivity analyses are consistent with the results of the OS primary analysis.~~

Reference ID: 4090305 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------XIAOPING JIANG 04/27/2017

~~LISA R RODRIGUEZ 04/27/2017~~

~~KUN HE 04/27/2017~~

~~Reference ID: 4090305 U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research Office of Translational Sciences Office of Biostatistics~~

~~S TATISTICAL R E V I E W A N D E VALUATION~~

~~CLINICAL STUDIES~~

NDA #: 203085 Supplement #: 007 Drug Name: Stivarga (Regorafenib) Indication: The treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4) Applicant: Bayer HealthCare Pharmaceuticals Received Dates: October 6, 2016; October 31, 2017 PDUFA Date: April 30, 2017 Review Type: Priority Biometrics Division: Division of Biometrics V Statistical Reviewer: Xiaoping (Janet) Jiang, Ph.D. Concurring Reviewers: Lisa Rodriguez, Ph.D., Team Leader Rajeshwari Sridhara, Ph.D., Division Director Medical Division: Division of Oncology Products 2 Clinical Team: Lorraine Pelosof, M.D., Ph.D., Clinical Reviewer Steven Lemery, M.D., Clinical Team Leader Patricia Keegan, M.D., Division Director Project Manager: Anuja Patel, M.P.H.

~~Keywords: Stratified log-rank test, Kaplan-Meier method, Cox regression~~

~~Reference ID: 4080784 Table of Contents 1 EXECUTIVE SUMMARY ...... 3~~

~~2 INTRODUCTION ...... 4~~

~~2.1 OVERVIEW...... 4 2.2 DATA SOURCES ...... 4 3 STATISTICAL EVALUATION ...... 4~~

3.1 DATA AND ANALYSIS QUALITY...... 4 3.2 EVALUATION OF EFFICACY...... 5 3.2.1 Study Design and Endpoints ...... 5 3.2.2 Statistical Methodologies...... 8 3.2.3 Patient Disposition, Demographic and Baseline Characteristics ...... 9 3.2.4 Results and Conclusions ...... 12 3.3 EVALUATION OF SAFETY...... 24 3.4 BENEFIT-RISK ASSESSMENT...... 24 4 FINDINGS IN SPECIAL/SUBGROUP POPULATIONS...... 25

~~4.1 GENDER, RACE, AGE, AND GEOGRAPHIC REGION...... 25 4.2 OTHER SPECIAL/SUBGROUP POPULATIONS ...... 25 5 SUMMARY AND CONCLUSIONS ...... 26~~

~~5.1 STATISTICAL ISSUES ...... 26 5.2 COLLECTIVE EVIDENCE...... 26 5.3 CONCLUSIONS AND RECOMMENDATIONS...... 27 5.4 LABELING RECOMMENDATIONS ...... 27~~

~~Reference ID: 4080784 1 EXECUTIVE SUMMARY~~

The applicant submitted an efficacy supplement of New Drug Application (sNDA) with results and data from Study 15982 (also known as “RESORCE”) entitled ‘A randomized, double blind, placebo-controlled, multicenter phase III study of regorafenib in patients with hepatocellular carcinoma (HCC) after sorafenib’ and other studies to seek an approval of regorafenib for the proposed indication: ‘treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4)

The primary endpoint was overall survival (OS). The primary analysis was a stratified log-rank test on the intent-to-treat population, consisting of all randomized patients. Based on 373 death events, the OS result demonstrated that the patients had statistically significant improvement in survival time when treated with regorafenib plus BSC compared to the patients treated with BSC alone (stratified log-rank p-value <0.0001). The estimated hazards ratio (HR) of OS was 0.63 with 95% confidence interval (CI) (0.50, 0.79) in favor of the treatment with regorafenib plus BSC. The estimated median OS was 10.6 months (95% CI: 9.1, 12.1) for patients in the regorafenib arm plus BSC versus 7.8 months (95% CI: 6.3, 8.8) for the patients in the placebo plus BSC arm.

The results of the secondary endpoint progression-free survival (PFS) assessed using modified Response Evaluation Criteria in Solid Tumors (mRECIST) also showed that treatment with regorafenib plus BSC statistically significantly delayed time to progression/death for patients compared to the treatment with BSC alone (stratified log-rank p-value <0.0001) with an estimated HR of 0.46 (95% CI: 0.37, 0.56). The results of PFS assessed using RECIST and the results of PFS assessed using mRECIST are similar (RECIST PFS: stratified log-rank p-value <0.0001 with an estimated HR of 0.43 and 95% CI: [0.35, 0.52]). The observed objective tumor response rate (ORR) per mRECIST (RECIST) was 10.6% (6.6%) in patients treated with regorafenib plus BSC versus 4.1% (2.6%) in patients treated with BSC alone. Whether to include results of PFS and ORR based on mRECIST or based on RECIST to the label of regorafenib is deferred to the clinical review team.

~~Whether the results from Study RESORCE provide a favorable benefit to risk ratio to support an approval of regorafenib for the proposed indication will be determined by the clinical review team.~~

~~Reference ID: 4080784 2 INTRODUCTION~~

~~2.1 Overview~~

Stivarga (regorafenib) is a bi-aryl urea that inhibits multiple kinases involved in tumor cell proliferation and neo-angiogenesis. On September 27, 2012, FDA approved Stivarga for “the treatment of patients with metastatic colorectal cancer (CRC) who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy, and, if KRAS wild type, an anti-EGFR therapy” On February 25, 2013, FDA approved Stivarga for the second indication ‘locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate.’ On October 6, 2016, the applicant submitted data and results from efficacy study 15982 (RESORCE) and human pharmacokinetic studies to seek an approval of regorafenib for a proposed indication for patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4) .

Study RESORCE was a phase 3, randomized, double-blind, placebo-controlled, and multicenter study of regorafenib in patients with hepatocellular carcinoma (HCC) after failed prior systemic treatment with sorafenib. The primary objective of the study was to evaluate efficacy and safety of regorafenib in patients with HCC in terms of overall survival (OS). Among 573 eligible patients, 379 patients were randomized to receive regorafenib plus BSC and 194 patients were randomized to receive match placebo plus BSC. The first visit of the first patient was on May 14, 2013 and the last visit of the last patient was on February 29, 2016. The primary completion date for the study was February 29, 2016. Study RESORCE was conducted at 152 study centers from Australia, the United States, and some countries in Europe and Asia.

~~The secondary study objectives included evaluating efficacy of regorafenib in terms of time to progression (TTP), progression-free survival (PFS) and objective tumor response rate (ORR).~~

~~2.2 Data Sources~~

~~Data used for this review were from the electronic submission received on October 6, 2016. The link was “\\CDSESUB1\evsprod\NDA203085\203085.enx”~~

~~3 STATISTICAL EVALUATION~~

~~This section focuses on efficacy evaluation for the efficacy study RESORCE.~~

~~3.1 Data and Analysis Quality~~

The quality of submitted data allowed this reviewer to reproduce the results of the primary analysis and other major submitted efficacy analyses. Also, the statistical analysis plan (SAP) was provided in the sNDA submission.

Reference ID: 4080784 Reviewer’s Comments: 1. During the review process, several information requests were sent to the applicant to request the SAS programs that were used to conduct major efficacy results and for clarification regarding the issues in the submitted SAS programs. The issues were resolved based on the applicant’s submitted responses.

~~3.2 Evaluation of Efficacy~~

~~3.2.1 Study Design and Endpoints~~

RESORCE was designed as a randomized, double-blind, placebo-controlled, multi-center phase III study to evaluate efficacy and safety of regorafenib in patients with HCC who had failed prior systemic treatment with sorafenib. The inclusion criterion included histological or cytological confirmation of HCC or non-invasive diagnosis of HCC per American Association for the Study of Liver Diseases (AASLD) criteria in patients with a confirmed diagnosis of cirrhosis; failure to prior treatment with sorafenib (defined as documented radiological progression according to the radiology charter); Barcelona Clinic Liver Cancer (BCLC) stage Category B or C that could not benefit from treatments of established efficacy with higher priority such as resection, local ablation, chemoembolization, or systemic sorafenib.

~~Eligible patients who progressed on sorafenib were randomized in a 2:1 ratio, using a telephone Interactive Voice Response System (IVRS), to receive one of the two following treatments:~~

Experimental Arm: 160 mg (4 x 40 mg tablets) every day; 3 weeks on/1 week off in each 4 week cycle plus BSC (best supportive care) Control Arm: Matching 4 x 40 mg placebo tablets every day; 3 weeks on/1 week off in each 4 week cycle plus BSC

The randomization was stratified by geographical region (Asia vs. rest of world (ROW)), ECOG performance status (0 vs. 1), AFP levels (<400 ng/mL vs. ≥400 ng/mL), extrahepatic disease (presence vs. absence), and macrovascular invasion (presence vs. absence).

~~Figure 3.1 shows the overall study design of RESORCE.~~

~~Reference ID: 4080784 Figure 3.1 Overall Study Design~~

~~[Source: Clinical Study Report Figure 7.1]~~

Per the protocol, the treatment continued until disease progression defined by modified Response Evaluation Criteria in Solid Tumors (mRECIST) for Hepatocellular Carcinoma, unacceptable treatment-related toxicity, or patient or physician decision to discontinue. The standard solid tumor response evaluation criterion RECIST 1.1 measures just the diameter of a lesion, irrespective of viability or necrosis. In contrast to RECIST, mRECIST does not include necrotic tissue into the measurement, only viable parts of tumor lesions. For more details regarding the difference of the two criteria, refer to Dr. Pelosof’s clinical review. All patients entered the follow-up period upon discontinuation of study treatments (regorafenib or placebo) for survival until death was documented, except for those patients who specifically withdrew consent to follow-up.

The primary endpoint was OS. OS was defined as the time from date of randomization to death due to any cause. OS for patients who were not known to have died were censored at their last date of being known alive or at the database cutoff date, whichever came first. The assessment of survival status was performed every month until death, consent withdrawal, lost to follow up, or study end.

Assuming that the true OS hazard ratio was 0.70 with 8 months in the placebo arm and 11.4 months in the regorafenib arm, a total of 370 events could provide 90% power to detect statistically significant difference in OS between two arms at a 2-sided significance level of 0.05. A sample size of 560 patients would provide 370 events in approximately 33 months of study

~~Reference ID: 4080784 duration, assumed accrual time was 22 months and accrual rate 25 patients/month ramp-up 3 months.~~

The secondary endpoints in the study included TTP, PFS and ORR. Per the protocol and the SAP, TTP was defined as the time from randomization to radiological or clinical disease progression. TTP for patients without radiological or clinical tumor progression at the time of analysis were censored at their last date of tumor evaluation. PFS was defined as the time from randomization to disease progression (radiological or clinical) or death due to any cause, if death occurs before progression is documented. For the patient who did not have radiological or clinical tumor progression or death at the time of analysis, PFS was censored at the last date of tumor evaluation.

Tumor measurements (via CT or MRI scans) were conducted every 6 weeks ± 7 days until PD. After 8 cycles of treatment these assessments should be done every 12 weeks ± 14 days. Tumor response and disease progression were evaluated based on RECIST 1.1 criteria and the mRECIST criteria for HCC regarding the definition of Progressive Disease.

The following censoring rules for PFS were specified in the SAP. . For the patient who did not have tumor assessments after baseline, and - did not die, PFS was censored at day 1 - died within  the first 12+1 weeks after randomization for patients who discontinued treatment prior to cycle8  the first 24+2 weeks after randomization for patients who discontinued treatment after to cycle8, death was not considered as a PFS event - died later than  12+1 weeks after randomization for patients who discontinued treatment prior to cycle8,  24+2 weeks after randomization for patients who discontinued treatment after to cycle8, death was not considered as a PFS event and PFS will be censored at day 1.

. For the patient who changed therapy to something other than the study medication prior to observing progression, PFS was censored at the date of the last scan performed prior to the change of therapy. . For the patient who had progression after 2 consecutive missed or non-evaluable assessments (i.e., up to cycle 8 progression later than the last evaluable scan + 12 weeks + 1 week, after cycles 8 the last evaluable scan + 24 weeks + 2 weeks), PFS was censored at the date of the last evaluable scan before the 2 missing assessments. . For the patient who discontinued or withdrew early from the study without documented progression, PFS was censored on the date of the last evaluable tumor assessment unless the patient died within - 12+1 weeks after the last evaluable assessment for subjects who discontinued treatment prior to cycle8,

~~Reference ID: 4080784 - 24+2 weeks after the last evaluable assessment for subjects who discontinued treatment after to cycle8 In these cases, death was considered a PFS event.~~

ORR was defined as the rate of patients with complete response (CR) or partial response (PR) over all randomized patients. Patients prematurely discontinuing without an assessment were considered non-responders.

In Study RESORCE, Health-Related Quality of Life (HRQoL) for patients was assessed as a tertiary objective. HRQoL was measured by using the instrument FACT-Hep (Functional Assessment of Cancer Therapy-Hepatobiliary) and the instrument EQ-5D (EuroQoL-5 Dimensions questionnaire) on the same visit schedule.

EQ-5D contains a descriptive system which measures 5 health dimensions: mobility, self-care, usual activity, pain/discomfort, and anxiety/depression. Each dimension contains 3 levels of response to reflect the degree of problems patients have experienced: no problem (level 1), some problems (level 2), and extreme problems (level 3). These five health dimensions are summarized into a single score, the EQ-5D index score according to EQ-5D scoring information. The EQ-5D index score ranges -0.59 to 1 with higher scores representing better health states. The EQ-5D also contains a visual analog scale (EQ-VAS), which records the respondents’ self- rated health status on a vertical graduated visual analogue scale ranging from 0 (worst imaginable health state) to 100 (best imaginable health state).

FACT-Hep consists of five subscales: (1) physical well-being (PWB); (2) social/family well- being (SWB); (3) emotional well-being (EWB); (4) functional well-being (FWB); and the hepatobiliary cancer subscale (HCS). The PWB, FWB, SWB and EWB are summed to form the FACT-General (FACT-G) total score. The FACT-G and HCS score are summed to form the FACT-Hep total score.

In Study RESORCE, the FACT-Hep and EQ-5D were both self-administrated by the patients at Cycle 1, Day 1, (i.e. baseline for patient report outcomes [PRO]), at every cycle, and end-of- study visit at the start of the visit before seeing the physician.

~~3.2.2 Statistical Methodologies~~

Per the SAP, the primary analysis of OS was a log-rank test stratified by five randomized stratified factors at one-sided significance level of 0.025. The primary analysis was performed according to treatment groups as randomized, with stratification strata as recorded in the IVRS. The Kaplan-Meier method was used to provide OS curves for each treatment arm. Hazard ratio and its 95% confidence interval were estimated using the Cox proportional hazards model stratified by the stratification factors that were used in the primary analysis of OS.

The same statistical analysis methods were applied to the secondary endpoints TTP and PFS analyses. In order to control type I error rate (alpha), a hierarchical order for testing the secondary endpoints TTP and PFS was specified in the SAP, i.e. after the primary analysis of OS

Reference ID: 4080784 shows statistically significant improvement in favor of Regorafenib, TTP was tested first and then PFS was tested. There were two planned OS interim analyses: when 30% and 75% of 370 required death events were observed. The first OS interim analysis was for futility pmposes. An O'Brien-Fleming alpha spending function approach was used to determine the significance thresholds based on the actual number of events by the time of the analyses.

ORR was compared between treatment groups using the Cochran-Mantel-Haenszel (CMH) test, adjusting for the same strntification factors as for the primaiy analysis of OS. Estimates and their 95% confidence intervals were computed for each ti·eatment group. The difference in ORR between the regorafenib and placebo group and the con esponding 95% confidence intervals were also calculated.

Per SAP, HRQoL measures as evaluated by EQ-5D and FACT-Hep instruments were analyzed as te1iiaiy endpoints. Descriptive statistics were provided for the FACT-Hep questionnaire (each domain score including the HCS and the FACT-Hep total score) and for the EQ-5D index score (obtained by summai·izing the five health dimensions) and visual analog scale score 01 AS) at each assessment time and for change from baseline by ti·eatment group.

~~Reviewer's Comments: 2. During the communication between FDA and the applicant befpre the sNDA submission, 4 FDA had conveyed to the sp_onsor that FDA didn't agree Cb>1~~

3. Per the protocol and SAP, disease progression in the PFS definition was assessed based on REC/ST 1.1 and the mRECIST, respective~y. However, SAP did not specify which criteria should be used as the primary analysis of PFS.

4. Per the SAP, ORR would be compared between treatment groups using the CMH test. Because testing ORR was not included in the pre-specified hierarchical test order for the secondary endpoints, the comparison should be considered exploratory analysis.

5. The data monitoring committee (DMC) recommended the study to be continued after reviewing the first interim analysis of OS. Also DMC recommended to remove the second interim ana~ysis of OS. The applicant followed the DMC recommendation as indicated in the protocol Amendment 4. According to O'Brien-Fleming spending function, a very small alpha (0. 00000696) was allocated to the first interim analysis.

~~3.2.3 Patient Disposition, Demographic and Baseline Characteristics~~

~~There were 843 patients screened to Study RESORCE. Among the 843 patients, 270 patients (32%) failed screening, and 573 patients (68%) were eligible to be randoinized to the study.~~

~~Reference ID: 4080784 Table 3.1 summarizes the disposition for all randomized patients as of August 05, 2016, the date for database locked for the clinical study report (CSR).~~

~~Table 3.1 Patient Disposition~~

~~Placebo + BSC Regorafenib + BSC~~

~~Randomized, n (%) 194 (100.0) 379 (100.0)~~

~~Never treated 1(0.5) 5 (1.3)~~

~~Started treatment 193(99.5) 374 (98.7)~~

~~Ongoing with treatment (as of LPLV) 10 (5.2) 65 (17.2)~~

~~Terminated treatment 183 (94.3) 309 (81.5)~~

~~Primary reason for terminating treatment~~

~~Progression disease 131 (68) 170 (45)~~

~~Withdrawal by subject 5 (2.6) 26 (6.9)~~

~~AE associated with clinical disease progression 28 (14.4) 56 (14.8)~~

~~AE not associated with clinical disease progression 12 (6.2) 47 (12.4)~~

~~Death 0 5 (1.3)~~

~~Other 5 (2.6) 2 (0.5) [Source: Clinical Study Report Table 8-1]~~

~~Reviewer’s Comments: 6. Note that there are more patients who withdrew treatment by themselves in regorafenib arm (6.9%) than placebo arm (2.6%).~~

~~Table 3.2 summarizes this reviewer’s results of the patients’ demographics.~~

Reference ID: 4080784 Table 3.2 Demographics for All Randomized Patients Placebo + BSC (n=194) Regorafenib + BSC (n=379) Total (n=573) Gender, n (%) Male 171 (88.1) 333 (87.9) 504 (88.0) Female 23 (11.9) 46 (12.1) 69 (12.0) Age, year Median (range) 62 (23 - 83) 64(19 - 85) 63 (19 - 85) Age Group n (%) <65 116 (59.8) 199 (52.5) 315 (55.0) >=65 78 (40.2) 180 (47.5) 258 (45.0) Region Asia 73 (37.6) 143 (37.7) 216 (37.7) Rest of the world 121 (62.4) 236 (62.3) 357 (62.3) Race/ethnic group, n (%) White 68 (35.1) 138 (36.4) 206 (36.0) Asian 78 (40.2) 156 (41.2) 234 (40.8) Other 3 (1.5) 8 (2.1) 11 (1.9) Not reported 45 (23.2) 77 (20.3) 122 (21.3)

~~Some major baseline disease characteristics are summarized in Table 3.3.~~

Table 3.3 Baseline Disease Characteristics for All Randomized Patients Placebo + BSC (n=194) Regorafenib + BSC (n=379) ECOG performance status (CRF*), n (%) 0 130 (67.0) 247 (65.2) 1 64 (33.0) 132 (34.8) Etiology of HCC, n (%) Hepatitis B 73 (37.6) 143 (37.7) Hepatitis C 41 (21.1) 78 (20.6) Alcohol use 55 (28.4) 90 (23.7) Genetic / Metabolic 6 (3.1) 16 (4.2) Non-Alcoholic steatohepatitis (NASH) 13 (6.7) 25 (6.6) Unknown 32 (16.5) 66 (17.4) Other 4 (2.1) 12 (3.2) BCLC stage at study entry, n (%) A (Early Stage) 0 1 (0.3) B (Intermediate Stage) 22 (11.3) 53 (14.0) C (Advanced Stage) 172 (88.7) 325 (85.8) Child-Pugh score, n (%) 5 118 (60.8) 244 (64.4) 6 70 (36.1) 129 (34.0) 7 5 (2.6) 5 (1.3) 8 or missing 1 (0.5) 1 (0.3) [Source: Clinical Study Report Table 8-5] *based on data collected in case report form (CRF) 11

~~Reference ID: 4080784 Reviewer’s Comments: 7. As shown in Table 3.2 and Table 3.3, the demographic and major baseline disease characteristics appear balanced between the two treatment arms.~~

~~3.2.4 Results and Conclusions~~

3.2.4.1 Results of Primary Endpoint As August 05, 2016, a total of 373 death events were observed, 3 more events than the requirement for conducting the final overall survival analysis. Table 3.4 summaries the applicant’s analyses of OS.

~~Table 3.4 Primary Analysis of Overall Survival~~

~~Placebo + BSC Regorafenib + BSC n=194 n=379~~

~~Number of Event (%) 140 (72.2) 233 ( 61.5)~~

~~Median OS in Months (95% CI) 7.8 (6.3, 8.8) 10.6 (9.1, 12.1)~~

~~Hazard Ratio (95%CI) 0.63 (0.50, 0.79)~~

P-valuea <0.0001 alog-rank test stratified by line of therapy (3rd-line vs. 4th-line or beyond), geographical region (Asia vs. rest of world (ROW)), ECOG-PS (0 vs. 1), Alpha-fetoprotein level (<400 ng/mL vs. ≥400 ng/mL), presence vs. absence of extrahepatic disease, and presence vs. absence of macrovascular invasion.

Reviewer’s Comments: 8. This reviewer has verified the applicant’s OS result shown in Table 3.4. Taking account a very small alpha (0.00000696 according to O’Brien-Fleming spending function) that was spent on the first interim analysis, the alpha available for the primary analysis was approximately 0.05. The OS primary analysis result in Table 3.4 demonstrated that the patients treated with regorafenib plus BSC had statistically significant improvement in survival time compared to the patients treated with placebo plus BSC.

~~Figure 3.2 displays the reviewer’s Kaplan-Meier curves of OS.~~

~~Reference ID: 4080784 Figure 3.2 Kaplan-Meier curves of OS~~

Reviewer’s Comments: 9. The primary analysis of OS shown in Table 3.4 was conducted using stratification values as recorded in the IVRS. The applicant also conducted sensitivity analyses such as stratified analysis using strata recorded in case report form (CRF) and unstratified analysis to evaluate the robustness of the OS primary analysis result. Table 3.5 summarizes the sensitivity analyses.

Table 3.5 Sensitivity Analyses of Overall Survival Placebo + BSC Regorafenib + BSC n=194 n=379 Number of Event (%) 140 (72.2) 233 ( 61.5) Median OS in Months (95%CI) 7.8 (6.3, 8.8) 10.6 (9.1, 12.1) Stratified Analysis using strata recorded from CRF Hazard Ratio (95%CI) 0.66 (0.53, 0.83) P-valuea 0.0003 Unstratified analysis Hazard Ratio(95%CI) 0.67 (0.55, 0.83) P-valuea 0.0002 alog-rank test stratified by line of therapy (3rd-line vs. 4th-line or beyond), geographical region (Asia vs. rest of world (ROW)), ECOG-PS (0 vs. 1), Alpha-fetoprotein level (<400 ng/mL vs. ≥400 ng/mL), presence vs. absence of extrahepatic disease, and presence vs. absence of macrovascular invasion.

Reference ID: 4080784 Reviewer’s Comments: 10. This reviewer has verified the two sensitivity analyses. As shown in Table 3.5, the results of OS sensitivity analyses are consistent with the result of the OS primary analysis.

3.2.4.2 Results of Secondary Endpoints Progression-free survival (PFS) and objective response rate (ORR) were the secondary endpoints evaluated in the study. Table 3.6 summaries the applicant’s and this reviewer’s PFS analyses based on mRECIST and RECIST, respectively.

~~Table 3.6 Analyses of Progression Free Survival~~

~~Placebo + BSC Regorafenib + BSC n=194 n=379~~

~~PFS assessed by using mRECIST~~

~~Number of Events (%) 181 (93.3) 293 (77.3) progression 173 274 death 8 19~~

~~Median PFS in months (95% CI) 1.5 (1.4, 1.6) 3.1 (2.8, 4.2)~~

~~Hazard Ratio (95% CI) 0.46 (0.37, 0.56)~~

~~P-valuea <0.0001~~

~~PFS assessed by using RECIST~~

~~Number of Events (%) 184 (94.8) 288 (76.0) progression 175 270 death 9 18~~

~~Median PFS in months (95% CI) 1.5 (1.4, 1.5) 3.4 (2.9, 4.2)~~

~~Hazard Ratio (95% CI) 0.43 (0.35, 0.52)~~

P-valuea <0.0001 [Source: Clinical Study Report Table 9-1] aLog-rank test stratified by line of therapy (3rd-line vs. 4th-line or beyond), geographical region (Asia vs. rest of world (ROW)), ECOG-PS (0 vs. 1), Alpha-fetoprotein level (<400 ng/mL vs. ≥400 ng/mL), presence vs. absence of extrahepatic disease, and presence vs. absence of macrovascular invasion.

~~Figure 3.3 displays Kaplan-Meier curves of PFS assessed using mRECIST.~~

~~Reference ID: 4080784 Figure 3.3 Kaplan-Meier curves of PFS (per mRECIST)~~

~~Figure 3.4 displays Kaplan-Meier curves of PFS assessed using RECIST.~~

~~Figure 3.4 Kaplan-Meier curves of PFS (per RECIST)~~

~~Reference ID: 4080784 Reviewer's Comments: 11. As mentioned in one of the previous reviewer's comments, FDA had conve ed to the {li)\4 a7:l.£.licant that FDA didn't a ree~~

~~12. As shown in Table 3.6 , treatment with regorafenib plus BSC statistically significant~y delayed time to progression/death compared to treatment with BSC alone.~~

13. The SAP did not specify which criteria (RECIST or mRECIST) would be used for the primary ana~ysis of PFS. It is deferred to the clinical review team to determine which PFS result (based on mRECIST or RECIST) should be included to the product label.

14. There are 12 patients who had different status of progression disease (PD) determined by using mRECIST. Among these 12 patients, 7 patients had PD determined by mRECIST and did not have PD determined by RECIST. The PFS using RECIST for those patients was censored at the date of the first PD assessed by using mRECIST which is the date of last evaluable tumor assessment before treatment change or before death. Table 3. 7 summaries this reviewer's concordance analysis ofPD assessment using mRECIST or RECIST.

~~Table 3.7 Concordance of mRECIST and RECIST~~

~~Using RECIST, n (%) Placebo + BSC Regorafenib + Placebo Using mRECSIT, n (%) n=194 n=379~~

~~PD No PD PD No PD 267 PD 173 (89.2) 0 (0.0) (70.4) 7 (1.8)~~

~~No PD 3 (1.5) IO (5.2) 2 (0.5) 84 (22.2)~~

~~Reviewer's Comments: 15. As shown in Table 3. 7, the concordance rate of using mRECIST and using RECIST in placebo arm (94. 4%) is similar as the one (92. 6%) in regorafenib arm.~~

~~Table 3.8 summarizes the applicant's and this reviewer's ORR analyses based on ORR data using mRECIST and RECIST, respectively.~~

~~Reference ID: 4080784 Table 3.8 Results of Objective Response Rate and Duration of Response Placebo + BSC Regorafenib + BSC n=194 n=379~~

Using mRECIST Response rate (95%CI) 4.1% (1.8%, 8.0%) 10.6% (7.6%, 14.1%) Responders (CR+PR) 8 40 Complete response 0 2 Partial response 8 38 p-value (CMHa) 0.004728 Median of Duration of Response (months) 2.7 (1.9, NEb) 3.5 (1.9, 4.5)

Using RECIST Response rate (95%CI) 2.6% (0.8%, 5.9%) 6.6% (4.3%, 9.6%) Responders (CR+PR) 8 25 Complete response 0 0 Partial response 8 25 p-value (CMHa) 0.019991 Median of Duration of Response (months) 5.6 (2.3, NEb) 5.9 (1.4, 8.4) aCMH=Cochran–Mantel–Haenszel test. bNE=not estimable due to few events.

Reviewer’s Comments: 16. Because the secondary endpoint ORR was not included in the pre-specified hierarchical order for testing the secondary endpoints, the p-values in the ORR results should be considered exploratory. Also, because the SAP did not specify whether ORR data using mRECIST or using RECIST should be used in the ORR primary analysis, it is deferred to the clinical review team to determine which ORR result should be included in the product label. Higher response rates are observed in bother arms using mRECIST correspond to using RECIST.

~~3.2.4.3 Results of Health-related Quality of Life (HRQL) As a tertiary objective of the study, Health-Related Quality of Life (HRQoL) was evaluated by using instruments FACT-Hep and the EQ-5D.~~

EQ-5D During the treatment, among the patients who were expected to report to EQ-5D questionnaires in each visit (cycle) and the end of treatment visit, 95% versus 78% of the patients provided the questionnaires at the visit through all cycles versus the end of treatment visit. More than 93% versus 77% of the expected questionnaires were all answered through all cycles versus the end of treatment visit. Table 3.9 shows this reviewer’s extract from the applicant’s summary of EQ-5D questionnaire completion analysis through cycle 10 and the end of treatment visit.

~~Reference ID: 4080784 Table 3.9: EQ-5D Questionnaire Completion Rate by Visit~~

#Patients Expected to Answer Questionnaires Completion Rate*

Visit Placebo + BSC Regorafenib + BSC Placebo + BSC Regorafenib + BSC Cycle 1 193 376 95.3% 97.1% Cycle2 176 340 95.5% 96.8% Cycle 3 97 269 96.9% 94.8% Cycle 4 66 220 98.5% 97.3% Cycle 5 51 184 96.1% 98.4% Cycle 6 34 157 97.1% 96.8% Cycle 7 29 128 86.2% 95.3% Cycle 8 19 116 100.0% 94.8% Cycle 9 14 93 93.3% 98.9% Cycle 10 11 85 90.9% 98.8% End of Treatment 144 236 79.2% 76.7% [Source: Clinical Study Report Table 14.2.4 / 1] *Completion Rate=#patients who answered all questions/#patients expected to answer questionnaires.

~~Table 3.10 summarizes the applicant’s descriptive statistics and changes from baseline for the EQ-5D index and EQ-VAS scores.~~

~~Reference ID: 4080784 Table 3.10: Mean EQ-5D index and VAS scores through Cycle 16~~

~~[Source: Clinical Study Report Table 9-16] a first unscheduled Abbreviations: C = cycle number; EOT = End of Treatment; StD = standard deviation;~~

~~Figures 3.5 displays mean scores with 95% CIs for the EQ-5D index score by treatment arm over the course of the study.~~

~~Reference ID: 4080784 Figure 3.5 Mean Scores of EQ-5D index with 95% CIs by Treatment over the Study~~

~~[Source: Clinical Study Report Figure 9-13]~~

~~Figures 3.6 displays mean scores with 95% CIs for the EQ-5D VAS by treatment arm over the course of the study.~~

~~Reference ID: 4080784 Figure 3.6 Mean Scores of EQ-5D VAS with 95% CIs by Treatment over the Study~~

~~[Source: Clinical Study Report Figure 9-14]~~

~~Reviewer’s Comments: 17. The analyses of EQ-5D and FACT-Hep are exploratory descriptive analyses and no inference should be drawn.~~

FACT-Hep Among the patients who were expected to provide FACT-Hep questionnaires in each visit (cycle) and the end of treatment visit, 95% versus 80% provided the questionnaires at the visit through all cycles versus the end of treatment visit. More than 86% versus 70% of the questionnaires were all answered through all cycles versus the end of treatment visit. Table 3.11

~~Reference ID: 4080784 shows this reviewer’s extract from the applicant’s summary of FACT-Hep questionnaire completion analysis through cycle 10 and the end of treatment visit.~~

~~Table 3.11: FACT--Hep Questionnaire Completion Rate by Visit~~

#Patients Expected to Answer Questionnaires Completion Rate*

Visit Placebo + BSC Regorafenib + BSC Placebo + BSC Regorafenib + BSC Cycle 1 193 376 87.6% 88.8% Cycle2 176 340 89.2% 87.4% Cycle 3 97 269 90.7% 87.7% Cycle 4 66 220 89.4% 86.8% Cycle 5 51 184 86.3% 89.7% Cycle 6 34 157 82.4% 93.0% Cycle 7 29 128 79.3% 89.8% Cycle 8 19 116 89.5% 90.5% Cycle 9 15 94 80.0% 90.4% Cycle 10 11 85 81.8% 90.6% End of Treatment 144 236 71.5% 70.3% [Source: Clinical Study Report Table 14.2.4 / 10] *Completion Rate=#patients who answered all questions/#patients expected to answer questionnaires.

~~Table 3.12 summarizes the applicant’s descriptive statistics and changes from baseline for the FACT-Hep total scores through Cycle 16.~~

~~Reference ID: 4080784 Table 3.12: FACT-Hep Total Score and Change from Baseline through Cycle 16~~

~~[Source: Clinical Study Report Table 9-17]~~

~~Figure 3.7 displays mean FACT-Hep total scores with 95% CIs by treatment arm over the course of the study.~~

~~Reference ID: 4080784 Figure 3.7: FACT-Hep Mean with 95% CIs by Treatment over the Study~~

~~[Source: Clinical Study Report Figure 9-15]~~

~~3.3 Evaluation of Safety Please refer to Dr. Pelosof’s clinical review for safety evaluation of regorafenib.~~

3.4 Benefit-Risk Assessment The results of Study RESORCE show statistically significant improvement in overall survival and reducing time to progression/death in patients with HCC who had failed prior systemic treatment with sorafenib when treated with regorafenib plus BSC instead of BSC alone.

Reference ID: 4080784 Whether the results from RESORCE provide a favorable benefit to risk ratio to support an approval of regorafenib for the proposed indication is deferred to the clinical review team.

~~4 FINDINGS IN SPECIAL/SUBGROUP POPULATIONS~~

4.1 Gender, Race, Age, and Geographic Region This reviewer has conducted OS analyses in the subgroups defined by age (greater than 65, less than or equal to 65 years), gender (male, female), and region (Asia vs. Rest of World; US vs. non-US). Table 4.1 summarizes this reviewer’s OS analyses in the demographic subgroups.

Table 4.1: OS Analyses in Demographics Subgroups Subgroup #Events/#Patients HR* (95%CI*) Age <65 205/305 0.65 (0.49, 0.87) >=65 168/258 0.74 (0.54, 1.02) Gender Male 327/504 0.65 (0.52, 0.82) Female 69/89 0.88 (0.48, 1.61) Region Asia 142/216 0.65 (0.46, 0.92) Rest of World 231/357 0.68 (0.52, 0.90) US 22/31 0.90 (0.35, 2.33) Non_US 351/542 0.67 (0.54, 0.83) *Hazard ratio and CIs are based on an unstratified Cox Regression Model.

~~Reviewer’s Comments: 18. As shown in the Table 4.1, there was no outlier subgroup observed. The subgroup analyses results are considered exploratory.~~

4.2 Other Special/Subgroup Populations This reviewer has conducted OS analyses in the subgroups defined by major baseline disease characteristics. The major baseline characteristics used to define the subgroups are ECOG performance status (0 vs. 1) at baseline, alpha fetoprotein (<400 ng/mL vs. ≥400 ng/mL), extrahepatic disease (presence vs. absence), macrovascular invasion (presence vs. absence), and etiology (Hepatitis B, Hepatitis C, alcohol use) based on the CRF. Table 4.2 summarizes this reviewer’s OS analyses in the subgroups based on major baseline characteristics.

Reference ID: 4080784 Table 4.2: OS Analyses in Baseline Characteristics Subgroups Subgroup #Events/#Patients HR* (95%CI*) ECOG Performance Status 0 231/377 0.61 (0.47, 0.80) 1 142/196 0.78 (0.55, 1.11) Alpha-Fetoprotein (CRF) <400ngml 194/324 0.67 (0.50, 0.90) >=400ngml 179/249 0.68 (0.50, 0.92) Extrahepatic Disease (CRF) absent 103/161 0.97 (0.63, 1.48) present 270/412 0.61 (0.47, 0.77) Macrovascular Invasion (CRF) absent 259/409 0.67 (0.52, 0.86) present 114/164 0.67 (0.46, 0.98) Etiology Hepatitis B No 238/357 0.73 (0.56, 0.95) Yes 135/216 0.58 (0.41, 0.82) Etiology Hepatitis C No 295/454 0.65 (0.51, 0.82) Yes 78/119 0.79 (0.49, 1.26) Etiology Alcohol Use No 273/428 0.59 (0.46, 0.76) Yes 100/145 0.92 (0.61, 1.38) Child-Pugh Status A5 222/362 0.59 (0.46, 0.79) A6 141/199 0.80 (0.57, 1.13) * Hazard ratio and CIs are based on an unstratified Cox Regression Model.

~~Reviewer’s Comments: 19. As shown in the Table 4.2, there was no outlier subgroup observed. The results of the baseline disease characteristic subgroup analyses are considered exploratory.~~

~~5 SUMMARY AND CONCLUSIONS~~

~~5.1 Statistical Issues This reviewer found no major statistical issue that impacted the overall conclusions.~~

5.2 Collective Evidence Based on the data from Study RESORCE, the result of the OS primary analysis demonstrated that patients with hepatocellular carcinoma (HCC) after treatment with sorafenib had statistically significant improvement in survival time when treated with regorafenib plus BSC instead of BSC alone (stratified log-rank p-value <0.0001) with an estimated HR of 0.63 (95% CI 0.50, 0.79). The estimated median survival time was 10.6 months (95% CI: 9.1, 12.1) for the patients treated with regorafenib plus BSC versus 7.8 months (95% CI: 6.3, 8.8) for the patients treated with

Reference ID: 4080784 BSC alone. The result of the secondaiy endpoint PFS assessed using mRECIST also showed that treatment with regorafenib plus BSC statistically significantly delayed time to progression /death (strntified log-rank p-value <0.0001) with an estimated HR of 0.46 (95% CI 0.37, 0.56) compai·ed to ti-eatment of BSC alone. The estimated median PFS assessed using mRECIST was 3.1 months (95% CI: 2.8, 4.2) for the patients treated with regorafenib plus BSC versus 1.5 months (95% CI: 1.4, 1.6) for the patients treated with BSC alone. The result of PFS assessed using RECIST is similar to the result of PFS assessed using mRECIST. The results of another secondaiy endpoint ORR are suppo1iive with the observed ORR per mRECIST (RECIST) of 10.6% (6.6%) in patients treated with regorafenib plus BSC versus 4.1% (2.6%) in patients treated with BSC alone.

5.3 Conclusions and Recommendations Based on analyses of OS and pre-specified key secondaiy endpoints from Study RESORCE, this reviewer concludes that treatment with regorafenib plus BSC statistically significantly prolongs survival time and delays time to progression/death for patients with HCC after treatment with sorafenib compared to the treatment with BSC alone. Whether the results from RESORCE provide a favorable benefit to risk ratio to suppo1i an approval of regorafenib for the proposed indication will be determined by the clinical review team.

5.4 Labeling Recommendations This reviewer recommends result of the OS primaiy analysis to be included in the product label of regorafenib. Whether to include results of PFS and ORR based on mRECIST or based on RECIST in the label of re orafenib is defened to the clinical review team. Ml"

Reference ID: 4080784 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------XIAOPING JIANG 04/06/2017

~~LISA R RODRIGUEZ 04/06/2017~~

~~RAJESHWARI SRIDHARA 04/07/2017~~

~~Reference ID: 4080784 CENTER FOR DRUG EVALUATION AND RESEARCH~~

~~APPLICATION NUMBER:~~

~~203085Orig1s007~~

~~CLINICAL PHARMACOLOGY AND BIOPHARMACEUTICS REVIEW(S) Office of Clinical Pharmacology Review (Amendment)~~

NDA or BLA Number NDA 203085 Link to EDR \\CDSESUB1\evsprod\NDA203085 Submission Date October 30, 2016 Submission Type Priority Brand Name Stivarga Generic Name Regorafenib Dosage Form and Strength 40 mg tablets Route of Administration Oral Proposed Indication Treatment of patients with Hepatocellular carcinoma (HCC) who have been previously treated with (b) (4) Applicant Bayer Associated IND IND 075642, (b) (4) OCP Review Team Youwei Bi, Ph.D.; Vadryn Pierre, Ph.D.; Jiang Liu, Ph.D.; Jeanne Fourie Zirkelbach, Ph.D. OCP Final Signatory

This review is an amendment to the OCP review for efficacy supplement of NDA 203085 regorafenib for the treatment of patients with hepatocellular carcinoma (HCC) which was submitted into DARRTS on April 6th, 2017. The section below is an amended version which corrects the number of patients with various degrees of hepatic impairment in section 2.4 Summary of Labeling Recommendations in the original OCP review.

~~Reference ID: 4098228 2.4 Summary of Labeling Recommendations (Amended)~~

~~The office of Clinical Pharmacology recommends the following labeling language with regard to hepatic impairment in section 12:~~

“Based on a population pharmacokinetic analysis, no clinically important differences in the mean total exposure of regorafenib, including M-2 and M-5, were noted amongst patients with normal liver function (total bilirubin and AST ≤ ULN, n=744), mild hepatic impairment (total bilirubin ≤ ULN and AST >ULN or total bilirubin >ULN to ≤1.5x ULN, n=437), and moderate hepatic impairment (total bilirubin >1.5x to ≤3x ULN and any AST, n=36). The pooled analysis included 391 patients with HCC of whom 116, 249, and 26 were categorized as having normal liver function, mild, and moderate hepatic impairment, respectively. The pharmacokinetics of regorafenib were not evaluated in patients with severe hepatic impairment (total bilirubin >3x ULN).”

Reference ID: 4098228 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------YOUWEI N BI 05/15/2017

~~VADRYN PIERRE 05/15/2017~~

~~JIANG LIU 05/15/2017~~

~~JEANNE FOURIE ZIRKELBACH 05/16/2017~~

~~Reference ID: 4098228 Office of Clinical Pharmacology Review~~

~~Reference ID: 4080727 Table of Contents~~

1. EXECUTIVE SUMMARY ...... 4 1.1 Recommendations ...... 4 1.2 Post-Marketing Requirements and Commitments ...... 4 2. SUMMARY OF CLINICAL PHARMACOLOGY ASSESSMENT ...... 5 2.1 Pharmacology and Clinical Pharmacokinetics ...... 5 2.2 Dosing and Therapeutic Individualization ...... 5 2.2.1 General dosing ...... 5 2.2.2 Therapeutic individualization ...... 5 2.3 Outstanding Issues ...... 5 2.4 Summary of Labeling Recommendations ...... 6 3. COMPREHENSIVE CLINICAL PHARMACOLOGY REVIEW ...... 6 3.1 Clinical Pharmacology Review Questions ...... 7 3.1.1 To what extent does the available clinical pharmacology information provide pivotal or supportive evidence of effectiveness? ...... 7 3.1.2 Is the proposed dosing regimen appropriate for the general patient population for which the indication is being sought? ...... 8 3.1.3 Is an alternative dosing regimen and/or management strategy required for subpopulations based on intrinsic factors? ...... 10 4. APPENDICES ...... 13 4.1 Summary of Bioanalytical Method Validation and Performance ...... 13 4.2 Clinical PK and/or PD Assessments ...... 13 4.3 Population PK and/or PD Analyses ...... 13 4.4 Exposure-Response Analyses ...... 28

Reference ID: 4080727 1. EXECUTIVE SUMMARY The applicant submitted an efficacy supplement to support the approval of regorafenib for the treatment of patients with hepatocellular carcinoma (HCC) who have been previously treated with (b) (4)

To support the efficacy of regorafenib in the proposed HCC indication the applicant submitted results from trial 15982, entitled: “A randomized, double-blind, placebo-controlled, multicenter study to evaluate efficacy and safety of regorafenib in HCC patients who had previously failed sorafenib”. The trial included 573 subjects stratified by geographical region [Asia, ROW (rest of the world)], ECOG status [0 vs 1], AFP levels (<400 ng/mL vs ≥400 ng/mL), presence of extrahepatic disease, and presence of macrovascular invasion). Patients were randomized in a 2:1 ratio to treatment with regorafenib (n=379) or placebo (n=194). The study met its primary endpoint as demonstrated by a statically significant improvement in median overall survival (OS) for patients treated with regorafenib, 10.1 months (95% CI: 9.1, 12.1) compared to those treated with placebo 7.8 months (95% CI: 6.3, 8.3); [Hazard Ratio (HR) 0.63 (95% CI: 0.50, 0.79) (p-value=0.000020)]. Notably, the proposed dosage regimen for the proposed indication is acceptable based on the clinical pharmacology properties of regorafenib and efficacy data from trial 15982. Based on the population pharmacokinetic post-hoc analysis, no clinically important differences in the mean total exposure of regorafenib, including M2 and M5, were noted amongst patients with mild tomoderate hepatic impairment, or those with normal liver function according to NCI criteria for organ impairment. Regorafenib is not recommended for use in patients with severe hepatic impairment defined by total bilirubin >3x ULN or Child-Pugh C as it has not been studied in this population. No conclusive exposure-response relationships were identified between regorafenib exposure and relevant efficacy and safety data.

1.1 Recommendations The Office of Clinical Pharmacology Divisions of Clinical Pharmacology V and Pharmacometrics have reviewed the information contained in this supplement for NDA 203085. This supplement is approvable from a clinical pharmacology perspective. The key review issues with specific recommendations/ comments are summarized below: Review Issue Recommendations and Comments

Pivotal or supportive evidence of The primary evidence of effectiveness comes from a randomized, effectiveness double-blind, placebo-controlled, multicenter trial (15982 [RESORCE]). General dosing instructions The recommended dose for regorafenib is 160 mg orally once daily with a light-fat meal for 21 consecutive days followed by 7 days off treatment in a complete cycle of 28 days. Dosing in patient subgroups No dose individualization is recommended based on intrinsic (intrinsic and extrinsic factors) factors. An appropriate dose has not been established for patients with severe hepatic impairment defined as total bilirubin >3x ULN or Child-Pugh C.

~~Reference ID: 4080727 Labeling Generally acceptable. The review team has specific content and formatting change recommendations.~~

~~Bridge between the to-be- Not applicable. To-be-marketed formulation was used in clinical marketed and clinical trial trials. formulations~~

~~Other (specify) Not applicable.~~

~~1.2 Post-Marketing Requirements and Commitments Not applicable~~

~~2. SUMMARY OF CLINICAL PHARMACOLOGY ASSESSMENT~~

2.1 Pharmacology and Clinical Pharmacokinetics Regorafenib is an oral multi-kinase inhibitor. An extensive summary of pharmacokinetics, relative bioavailability, food effect, dose-linearity, metabolism, excretion and drug interactions of regorafenib and its metabolites is provided in clinical pharmacology review for the original NDA 203085 submission (Reference ID: 3182650). In brief, regorafenib undergoes extensive and complex biotransformation. The primary metabolism pathway of regorafenib involves both CYP3A4 and UGT1A9. Following a single oral administration of radiolabeled regorafenib, approximately 47% (24% as metabolite) and 19% (17% as glucuronides) of the dose was excreted as unchanged regorafenib in feces and urine, respectively. The active metabolites of regorafenib, M2 and M5, have demonstrated in vitro pharmacologic activity similar to that of parent

~~regorafenib. The mean elimination T1/2 for regorafenib, M2, and M5 was 28, 25, and 51 hrs, respectively.~~

~~2.2 Dosing and Therapeutic Individualization~~

2.2.1 General dosing The recommended dosing regimen for regorafenib is 160 mg orally once daily with a light-fat meal for 21 consecutive days followed by 7 days off treatment in a complete cycle of 28 days. Regorafenib is available as 40 mg film-coated tablets.

~~2.2.2 Therapeutic individualization Hepatic Impairment~~

The pharmacokinetics of regorafenib, M-2, and M-5 were evaluated in 14 patients with hepatocellular carcinoma (HCC) and mild hepatic impairment (Child-Pugh A); 4 patients with HCC and moderate hepatic impairment (Child-Pugh B); and 10 patients with solid tumors and normal hepatic function after the administration of a single 100 mg dose of regorafenib. No clinically important differences in the mean

Reference ID: 4080727 exposure of regorafenib, M-2, or M-5 were observed in patients with mild or moderate hepatic impairment compared to the patients with normal hepatic function. Based on a population pharmacokinetic post-hoc analysis, no clinically important differences in the mean total exposure of regorafenib, including M2 and M5, were noted amongst patients with normal liver function (n=36), mild hepatic impairment (n=744), and moderate hepatic impairment (n=437) according to NCI criteria. The pharmacokinetics of regorafenib were not evaluated in patients with severe hepatic impairment defined as total bilirubin >3x ULN or Child-Pugh C.

~~Drug-Drug Interactions~~

~~In brief, the concomitant use of regorafenib with strong CYP3A4 inducers or strong CYP3A4 inhibitors should be avoided.~~

~~2.3 Outstanding Issues Not applicable~~

~~2.4 Summary of Labeling Recommendations The office of Clinical Pharmacology recommends the following labeling language with regard to hepatic impairment in section 12:~~

“Based on a population pharmacokinetic analysis, no clinically important differences in the mean total exposure of regorafenib, including M2 and M5, were noted amongst patients with normal liver function (total bilirubin and AST  ULN, n=(b) ), mild hepatic impairment (total bilirubin  ULN and AST >ULN, or (4) total bilirubin >ULN to ≤1.5x ULN, n=(b) (4) , and moderate hepatic impairment (total bilirubin >1.5x to ≤3x ULN, and any AST, n=(b) (4) ). The pooled analysis included (b) (4) patients with HCC of which (b) (4) and (b) (4) were categorized as having normal liver function, mild, and moderate hepatic impairment, respectively. The pharmacokinetics of regorafenib were not evaluated in patients with hepatic impairment severity (total bilirubin >3x ULN).”

~~3. COMPREHENSIVE CLINICAL PHARMACOLOGY REVIEW~~

~~3.1 Clinical Pharmacology Review Questions~~

3.1.1 To what extent does the available clinical pharmacology information provide pivotal or supportive evidence of effectiveness? To support the efficacy of regorafenib in the proposed HCC indication the applicant submitted results from trial 15982. A statically significant OS benefit was identified for regorafenib with a HR of 0.63 (95% CI: 0.50, 0.79) (P-value=0.000020). The median OS was 10.6 months (95% CI 9.1, 12.1) in the regorafenib group compared to 7.8 months (95% CI 6.3, 8.8) in the placebo group. In addition, a treatment effect in favor of the regorafenib group with respect to the secondary endpoint of progression-free survival (PFS) was also identified with a HR of 0.46 (95% CI: 0.37. 0.56) (p-value<0.0001) using mRECIST criteria.

~~Reference ID: 4080727~~

Exposure-response (E-R) relationships were explored between the selected exposure metric, average exposure of parent regorafenib in cycle 1 (CAVD28), and overall survival for study 15982. The univariate Kaplan Meier plot for the CAVD28 exposure quartiles and placebo for OS is shown in Figure 1. There appears to be a trend of shorter OS in patients with lower average exposure in cycle 1 (CAVD28) compared to patients with medium to high average exposure. Of note, the median overall survival in the low exposure group was numerically better when compared to patients receiving placebo. However, a strong association was not observed between OS and CAD28 in regorafenib-treated patients with a mortality HR equal to 0.984 (95%CI: 0.965, 1.004) based on a multivariate cox regression model adjusted for significant baseline covariates (ECOG status, baseline AFP value and baseline AST/ALT values). As a result, no evident and conclusive E-R relationship was identified between regorafenib exposure and overall survival.

~~Figure 1: (Univariate) Kaplan-Meier Curves for OS stratified by CAVD28 exposure quartiles and Placebo.~~

3.1.2 Is the proposed dosing regimen appropriate for the general patient population for which the indication is being sought? The proposed dosage of regorafenib is 160 mg once daily for 21 consecutive days followed by 7 days off treatment in a complete cycle of 28 days. As discussed in the section 3.3.1, according to trial 15982 the

~~Reference ID: 4080727 proposed dosage appears to be effective with acceptable safety in patients with HCC who have been previously treated with (b) (4) .~~

Dose selection for regorafenib in the HCC patients was based on Phase 1, 2, and 3 data for mCRC and GIST indications as well as the Phase 2 study in HCC patients (A51601). The proposed dosage, 160 mg regorafenib once daily for 21 consecutive days followed by 7 days off treatment in a complete cycle of 28 days, was approved for the treatment of mCRC in 2012 and GIST in 2013, respectively. Regorafenib exposures are comparable across different trials and different indications.

~~The geometric mean average exposure to free aggregate, sum of the free molar concentrations of regorafenib and its metabolites M2 and M5, during first 2 cycles was about 20 nM, which is higher than~~

the IC80 (16 nM) of regorafenib and M2 for VEGF receptors. The median exposure during cycle 1 in the st low exposure group (<1 quartile) of free aggregate (13.4 nM) is only slightly lower than IC80 but still well above the IC50 (4 nM). The observed exposure data for regorafenib and its two metabolites suggest that a dose of 160 mg once daily result in systemic exposure of regorafenib that is able to inhibit the VEGF receptors in in vitro studies.

Exposure-safety reltionships for regorafenib were evaluated for incidence rate of overall TEAEs, incidence rate of TEAE of special interest (Hypertension, Diarrhea, Fatigue and HFSR) and laboratory abnormalities (AST, ALT, bilirubin and platelet). Overall, no consistent and conclusive exposure- dependent relationships were identified between regorafenib exposure and the above mentioned safety data explored in the E-R analysis.

In summary, the proposed dosage, 160 mg regorafenib once daily for 21 consecutive days followed by 7 days off treatment in a complete cycle of 28 days, appears to be appropriate in patients with HCC who have been previously treated with (b) (4) based on the efficacy and safety data in a trial 15982. The selected dose was further supported by the evidence of effectiveness in the approved indications of mCRC and GIST. No evident and conclusive exposure-response relationships were identified for both efficacy and safety. However, there appears to be a trend for shorter OS in patients within the low exposure group.

3.2.3 Is an alternative dosing regimen and/or management strategy required for subpopulations based on intrinsic factors? A post-hoc population PK analysis showed that there are no clinically relevant effects of intrinsic factors on the systematic exposure of regorafenib and its metabolites. No dose adjustment based on intrinsic factors is needed.

Hepatic impairment: NCI criteria Among 1276 subjects from 16 clinical studies, 830, 413 and 33 subjects have normal, mild and moderate hepatic impairment based on NCI criteria [mild: total bilirubin less than or equal to upper limit of normal (ULN) and AST greater than ULN or total bilirubin between 1.0 to 1.5 times ULN and any AST, moderate: total bilirubin between 1.5 to 3 times ULN and any AST]. Overall, no influence of mild and moderate hepatic impairment at the start of treatment was observed on parent and total regorafenib exposure

~~Reference ID: 4080727 (Figure 2). The pharmacokinetics of regorafenib have not been studied in patients with severe hepatic impairment [severe: total bilirubin greater than 3 times ULN and any AST].~~

Figure 2: Comparison of Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) across Different Levels of Liver Impairment (NCI criteria) at Start of Treatment within RESORCE (Top) ,Phase III Studies (Middle) and all studies (Bottom).

~~(A) Parent nominal exposure (Cav,md,p) in RESORCE (B) Total nominal exposure (Cav,md,tot) in RESORCE~~

~~(C) Parent nominal exposure (Cav,md,p) in Phase III studies (D) Total nominal exposure (Cav,md,tot) in Phase III studies~~

~~(E) Parent nominal exposure (Cav,md,p) in all studies (F) Total nominal exposure (Cav,md,tot) in all studies~~

Reference ID: 4080727 Child-Pugh score The parent and total nominal exposures of subjects with Child-Pugh score level B are comparable to those with Child-Pugh score level A (Figure 3). However, this comparison is limited by the low number of subjects with Child-Pugh score B (n=1 in RESORCE study and n=5 in HCC studies). The pharmacokinetics of regorafenib have not been studied in patients with severe hepatic impairment (Child-Pugh C).

Figure 3: Comparison of Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) across Different Levels of the Child Pugh Score at Start of Treatment within RESORCE (Top) and All HCC Studies (Bottom).

~~(A) Parent nominal exposure (Cav,md,p) in RESORCE (B) Total nominal exposure (Cav,md,tot) in RESORCE~~

~~(C) Parent nominal exposure (Cav,md,p) in HCC studies (D) Total nominal exposure (Cav,md,tot) in HCC studies Source: Applicant’s population pharmacokinetic report, Page 51, Figure 7.2:16~~

Ethnic groups: In general, individual parent regorafenib and total exposures in patients with Asian race were comparable to exposures in patients with non-Asian race. No difference in exposure was observed among Japanese, Chinese and non-Asian patients. However, Asian patients being treated outside of Japan and China showed a trend of lower exposures compared to the other populations.

Other intrinsic factors: The influence of covariates on the parent nominal exposure of regorafenib (Cav,md,p) was further investigated using a stepwise generalized additive modeling approach. The following covariates were

Reference ID: 4080727 found to significantly influence regorafenib exposure: patient type, age, sex, BMI (body mass index), HB0 (baseline hemoglobin) and ALB0 (baseline albumin). However, the magnitude of the predicted changes in regorafenib exposure impacted by these significant covariates was considered small when compared to the overall observed variability, and no dose adjustment based on these covariates is needed.

~~Reference ID: 4080727 4. APPENDICES~~

4.1 Population PK and/or PD Analyses 4.1.1 Introduction Regorafenib is multi kinase inhibitor that inhibits tumor growth by suppressing tumor cell proliferation and tumor angiogenesis. Regorafenib has two active metabolites, M2 (BAY 75-7495) and M5 (BAY 81- 8752), which were shown to have similar pharmacologic activity to parent drug in in vitro and in vivo studies. In the previous submission for regorafenib, a population PK model was developed that described the concentration-time profiles of regorafenib, M-2, and M-5 in a pooled dataset of 14 sparsely and densely sampled Phase 1, 2 and 3 studies. This established integrated PK model was applied to the sparsely sampled PK data from the RESORCE study (trial 15982 for HCC indication) in order to calculate the individual exposure estimates, and conduct a covariate analysis of parent and total nominal exposure in the combined dataset of 16 clinical trials.

4.1.1.1 Established Population PK Model for Regorafenib and its Metabolites The final integrated model for regorafenib is characterized by first order absorption followed by a two compartmental disposition. M2 is formed in a concentration-dependent, non-linear manner from parent and follows a two-compartmental disposition. M5 is generated from M2 in a concentration-dependent, non-linear manner and also has a two-compartmental disposition. The structure for the final Pop-PK model is shown schematically in Figure 4. Figure 4: Schematic Representation of the PK Models of Regorafenib, M2 and M5.

~~Source: Sponsor’s Pop-PK report R-8931 page 77 Figure 8:1~~

~~Reference ID: 40807274092559 4.1.2 Application and Results~~

4.1.2.1 Estimation of individual exposure of regorafenib and its metabolites - 80% dataset The developed population PK model was applied to fit 2534 observations (846 parent, 828 M2 and 860 M5 observations, respectively) from 276 subjects available from an intermediate analysis dataset consisting of 80% of the total PK samples from Phase III study 15982 (RESORCE) in order to calculate the following individual exposure metrics:

a) Parent, M2, M5, and total average concentration over the first 24 hours after treatment start following actual dosing (Cav(0-24h),sd,p, Cav(0-24h),sd,m2, Cav(0-24h),sd,m5, Cav(0- 24h),sd,tot) b) Parent, M2, M5, and total average concentration over 24 h after 21 days of nominal dosing (160 mg QD) (Cav,md,p, Cav,md,m2, Cav,md,m5, Cav,md,tot) c) Parent, M-2, M-5, and total average concentration over the first 28 days after treatment start following actual dosing (3 weeks on/1 week off schedule) (Cav(0-28d),md,p, Cav(0-28d),md,m2, Cav(0-28d),md,m5, Cav(0-28d),md,tot) d) Parent, M-2, M-5, and total average concentration over the first 56 days after treatment start following actual dosing (3 weeks on/1 week off schedule) (Cav(0-56d),md,p, Cav(0-56d),md,m2, Cav(0-56d),md,m5, Cav(0-56d),md,tot)

Goodness-of-fit plots (Figure 5 - Figure 7) and prediction-corrected visual predicted check (Figure 8) showed the previously developed model adequately described the steady-state concentrations of regorafenib and its two metabolites, but bias was shown in the predictions of non-steady-state concentrations such as observations sampled in cycle 2 day 1. The impact of biased predictions at cycle 2 day 1 visit on the calculation of individual exposure metrics was evaluated by fitting the integrated model to the data including and excluding those observations. The comparison of the parent and total nominal exposure (Cav,md,p and Cav,md,tot) obtained including and excluding cycle 2 day 1 observations was shown in Figure 9. In general, the observations in cycle 2 day 1 had no influential impact on the estimates of parent and total nominal exposure.

Reviewer’s comments: The reviewer agrees that goodness-of-fit plots and simulation-based diagnostics (pc-VPC) showed that the prior integrated Pop-PK model adequately described the steady-state concentrations of regorafenib and its two metabolites M2 and M5, and a bias was shown in the predictions of non-steady-state concentrations (e.g. Cycle 2 Day 1). However, by comparing the parent and total nominal exposure (Cav,md,p and Cav,md,tot) obtained including and excluding cycle 2 day 1 observations, the reviewer agrees that these observations had no influential impact on the estimates of individual exposure metrics.

~~Reference ID: 4080727 Figure 5: Goodness-of-fit Plots of Regorafenib from Prior Pop-PK Model.~~

DV: Observations; PRED: Population Predictions; IPRED: Individual Predictions; CWRES: Conditional Weighted Residuals. Blue dots: steady-state concentrations. Red dots: non steady-state concentrations. Red solid line: Loess smooth through data.

~~Reference ID: 4080727 Figure 6: Goodness-of-fit Plots of Metabolite M2 from Prior Pop-PK Model.~~

~~Reference ID: 4080727 Figure 7: Goodness-of-fit Plots of Metabolite M5 from Prior Pop-PK Model.~~

~~Reference ID: 4080727 Figure 8: Prediction-Corrected Visual Predictive Check for Regorafenib, M2 and M5 on Nominal Day 1 and 15~~

~~Red solid lines are median percentiles for observed data. Black solid lines are median percentiles for simulated data. Green area is the 95% confidence interval (CI) around the simulated median.~~

~~Reference ID: 4080727 Figure 9: Comparison of Parent and Total Nominal Exposure (Cav,md,p and Cav,md,tot) obtained including and excluding cycle 2 day 1 visit observations.~~

~~4.1.2.2 Covariate analysis - 80% dataset~~

4.1.2.2.1 Graphical evaluation of individual exposure metrics The dataset used for the covariate analysis consisted of parent and total nominal exposure (Cav,md,p and Cav,md,tot) and the covariates of 1276 subjects from 16 clinical regorafenib trials. The graphical exploration of covariates and subpopulations with respect to parent and total nominal exposure is described below.

Study population Overall, the individual estimates of parent and total nominal exposure in study 15982 largely overlapped with those of the other studies including earlier Phase 3 studies (Figure 10). Among all studies there are no studies that stand out with clear higher than average exposure.

~~Reference ID: 4080727 Figure 10: Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) in All Studies.~~

White: phase I studies, grey: phase II studies, blue: phase III; Whiskers (Black bars): Range between the lowest observation still within 1.5 x the interquartile range (IQR) of the lower quartile and the highest observation still within 1.5 x interquartile range (IQR) of the upper quartile. Box: Range between lower and upper quartile. Notches: 95% confidence interval of the median. Black horizontal line: Median. Grey dots: Outliers (individual values outside of whiskers), blue dots: individual observations.

Hepatocellular cancer (HCC) The individual estimates of parent and total nominal exposure is comparable in patients with HCC across Phase 1, 2 and 3 studies (Figure 11). Individual parent and total exposures in patients with HCC are comparable to those from patients with colorectal cancer (CRC), while they are slightly higher than those in patients with gastrointestinal solid tumor (GIST) as previously reported (Figure 12). Healthy volunteers were found to have lower predicted exposures for parent and total nominal exposure than patients.

~~Figure 11: Distribution of Cav,md,p (Left) and Cav,md,tot (Right) in HCC Patients across Phase I, Phase II and Phase III studies.~~

~~Source: Applicant’s population pharmacokinetic report, Page 48, Figure 7.2:12~~

~~Reference ID: 4080727~~

~~Figure 12: Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) across patients with GIST, CRC, HCC and healthy volunteers.~~

Ethnic groups In general, individual parent and total exposures in patients with Asian race were comparable to exposures in patients with non-Asian race (Figure 13). No difference in exposure was observed among Japanese, Chinese and non-Asian patients (Figure 14). However, Asian patients being treated outside of Japan and China showed a trend of lower exposures compared to the other populations.

~~Figure 13: Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) in Asian and non-Asian Subjects of RESORCE (top) and Phase III Studies (Bottom).~~

~~(A) Parent nominal exposure (Cav,md,p) in RESORCE (B) Total nominal exposure (Cav,md,tot) in RESORCE~~

~~(C) Parent nominal exposure (Cav,md,p) in phase III studies (D) Total nominal exposure (Cav,md,tot) in phase III studies~~

Reference ID: 4080727 Source: Applicant’s population pharmacokinetic report, Page 48, Figure 7.2:13 Figure 14: Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) in Asian, Japanese, Chinese and Others of RESORCE (top) and phase III studies (bottom).

~~(A) Parent nominal exposure (Cav,md,p) in RESORCE (B) Total nominal exposure (Cav,md,tot) in RESORCE~~

~~(C) Parent nominal exposure (Cav,md,p) in phase III studies (D) Total nominal exposure (Cav,md,tot) in phase III studies Source: Applicant’s population pharmacokinetic report, Page 49, Figure 7.2:14~~

Liver impairment NCI Criteria Among 1276 subjects from 16 clinical studies, 830 have normal liver funcition, while 413 and 33 subjects have mild and moderate hepatic impairment, respectively, based on NCI criteria [mild: total bilirubin less than or equal to upper limit of normal (ULN) and AST greater than ULN or total bilirubin between 1.0 to 1.5 times ULN and any AST, moderate: total bilirubin between 1.5 to 3 times ULN and any AST]. Overall, no influence of mild and moderate hepatic impairment at the start of treatment was observed on parent and total regorafenib exposures (Figure 2). The pharmacokinetics of regorafenib have not been studied in patients with severe hepatic impairment [severe: total bilirubin greater than 3 times ULN and any AST].

Child-Pugh score The parent and total nominal exposure of subjects with Child-Pugh score level B is comparable to those with Child-Pugh score level A (Figure 3). However, this comparison is limited by the low number of subjects with Child-Pugh score B (n=1 in RESORCE study and n=5 in HCC studies). The pharmacokinetics of regorafenib have not been studied in patients with severe hepatic impairment (Child-Pugh C).

Reference ID: 4080727 4.1.2.2.2 Stepwise generalized additive modeling (GAM) The influence of covariates on the parent nominal exposures of regorafenib (Cav,md,p) were further investigated using a stepwise generalized additive modeling approach. Each potential covariate was first investigated by testing all the models of each individual covariate in the univariate manner. The covariates that resulted in decrease in AIC were carried forward to the stepwise multivariate GAM analysis. The backward deletion approach was then conducted to screen selected covariates. A covariate was kept in the multivariate model if its removal results in an increase in AIC. The following covariates were retained in the final statistical model: patient type, age, sex, BMI (body mass index), HB0 (baseline hemoglobin) and ALB0 (baseline albumin). The resulting parameter estimates are shown in Table 1. Similar modeling procedure was conducted to investigate the influence of covariates on the total nominal exposure (Cav,md,tot). The resulting parameter estimates are shown in Table 2.

~~Table 1: Parameter Estimates of the Covariate Model of the Log Transformed Parent Nominal Exposure (Cav,md,p)~~

~~Source: Applicant’s population pharmacokinetic report, Page 53, Table 7.2:13~~

~~Table 2: Parameter Estimates of the Covariate Model of the Log Transformed Average Concentration over 24h for Parent, M-2 and M-5 (Cav,md,tot)~~

~~Source: Applicant’s population pharmacokinetic report, Page 54, Table 7.2:14~~

Reference ID: 40807274092559 The clinical relevance of the impact of significant covariates on parent and total nominal exposure was assessed by computing the predicted difference between the regorafenib exposure at the 5th to 95th percentile of the covariate and the exposure at the median of the covariate for males and females separately. The predicted difference was then compared to the overall observed variability in the regorafenib exposure. Simulation results showed that each significant covariate can only explain a small fraction of the exposure variability (up to 11.4% for Cav,md,p and 15.7% for Cav,md,tot) (Figure 15). Therefore, the clinical relevance of the significant covariates on the regorafenib exposure was considered small, and no dose adjustment based on these covariates is needed.

~~Figure 15: Impact of the Continuous Covariates on the Variability of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (right) for Males and Females Separately.~~

~~(A) Impact of covariates on parent nominal exposure (Cav,md,p) (B) Impact of covariates on total nominal exposure (Cav,md,tot)~~

Range bars show difference between the 5th to 95th percentile and the median of Cav,md,p and Cav,md,tot in all cancer patients. Covariate bars show the predicted difference between the Cav,md,p at the 5th to 95th percentile of the covariate and the Cav,md,p at the median of the covariate in all cancer patients.

Reviewer’s comments: The reviewer agrees with the applicant in general. Graphical evaluation showed that parent or total nominal exposure appears to be comparable among patients with normal liver function, mild hepatic impairment and moderate hepatic impairment based on NCI criteria. The regorafenib exposure appears to be comparable between patients with Asian race and patients with non-Asian race. Stepwise generalized additive modeling has further identified the following statistically significant covariates on parent nominal exposure: patient type, age, sex, BMI (body mass index), HB0 (baseline hemoglobin) and ALB0 (baseline albumin). However, the impact of these covariates on the regorafenib exposure was small compared to the observed overall variability, and no dose adjustment based on these covariates is needed.

~~Reference ID: 4080727 4.1.2.3 Comparison of individual exposure of regorafenib between 80% and 100% dataset~~

4.1.2.3.1 Graphical Comparison After obtaining 100% full dataset for study 15982, the applicant further compared the regorafenib parent and total nominal exposure between the interim and final dataset. In the final dataset, individual parent and total nominal exposure metrics were obtained from 337 patients based on 3210 observations (1073, 1049 and 1088 for parent, M2 and M5, respectively) in total. As illustrated in Figure 16, the distribution of parent and total nominal exposure in the full and interim dataset (80%) of study 15982 were very similar.

~~Figure 16: Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) in RESORCE Study.~~

~~Red: based on dataset containing approximately 80% of RESORCE study data, Blue: based on dataset containing 100% of RESORCE study data.~~

~~Age~~

The parent and total nominal exposures were further compared graphically within different age categories; subjects < 65 years and ≥ 65 years and subjects < 65 years, 65-74 years, 75-84 years, and ≥ 85 years (Figure 17). The distribution of individual parent and total nominal exposures were comparable

~~Reference ID: 4080727 among these age categories, although slightly higher exposures were observed in the older age groups numerically.~~

~~Figure 17: Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) in Subjects within Different Age Categories.~~

~~(A) Parent nominal exposure (Cav,md,p) (B) Total nominal exposure (Cav,md,tot)~~

~~(C) Parent nominal exposure (Cav,md,p) (D) Total nominal exposure (Cav,md,tot) Alcohol Etiology~~

A possible effect of alcohol etiology (yes/no) on the regorafenib was also compared. As shown in the box plots in Figure 18, the distribution for individual parent and total nominal exposures of regorafenib were comparable in the univariate analysis of alcohol etiology.

~~Reference ID: 4080727 Figure 18: Distribution of Parent Nominal Exposure (Cav,md,p) (Left) and Total Nominal Exposure (Cav,md,tot) (Right) in the Alcohol Etiology Categories.~~

~~(A) Parent nominal exposure (Cav,md,p) (B) Total nominal exposure (Cav,md,tot)~~

~~4.2 Exposure-Response Analyses 4.2.1 Exposure-Response for Efficacy~~

The purpose of this exploratory exposure-response (E-R) analysis was to evaluate the relationship between individual exposures of parent or free aggregate regorafenib and overall survival (OS) based on the data from a phase 3 study 15982.

Six different exposure metrics were evaluated in the E-R analysis, which include the average parent regorafenib and free aggregate concentration over cycle 1 day 1, over 28 days in cycle 1 and over 56 days in cycle 1 and 2. The free aggregate concentration of regorafenib was calculated as the sum of the free/unbound molar concentrations of regorafenib and its pharmacologically active metabolites M2 and M5. The average exposure of parent regorafenib in cycle 1 (CAVD28) was selected for the multivariate Cox regression analysis as it shows the strongest association with OS (lowest AIC) in the univariate analysis. It also shows a wider range of exposure compared to average exposure in day 1, and has a higher number of observations compared to average exposure in first 2 cycles.

The univariate Kaplan Meier plot for the CAVD28 exposure quartiles and placebo for OS is shown in Figure 1. It appears a trend of shorter OS is observed in patients with lower average exposure in cycle 1 (CAVD28). After removing the pre-specified baseline covariates in a backward manner from full Cox regression model, ECOG status, baseline AFP value and AST/ALT baseline values were found to be significant covariates and retained in the final reduced model. The CAVD28 quartiles were found to be significantly associated with the overall survival compared to placebo (Table 3). However, the strong association was not observed between OS and CAD28 in regorafenib-treated patients using cox regression analysis adjusted for the same baseline covariates (HR=0.984; 95%CI: 0.965, 1.004) (Table 4). As a result, no evident and conclusive E-R relationship was identified between regorafenib exposure and OS.

~~Reference ID: 4080727 Table 3: Results of the Cox Regression Analysis of the Reduced Model for OS Including Placebo Data (n=514).~~

Covariate Category HR LLCI ULCI P-value of P-value of estimate category covariate CAVD28 Placebo 1 na na Na <0.001 Q1 0.784 0.574 1.072 0.128 Q2 0.55 0.393 0.768 <0.001 Q3 0.652 0.468 0.908 0.011 Q4 0.453 0.317 0.647 <0.001 ECOG stage 0 1 na na Na <0.001 >=1 1.577 1.253 1.985 <0.001 AFP level <400 ng/mL 1 na na Na 0 >=400 ng/mL 2.022 1.617 2.528 <0.001 Max of AST <= 1.5*ULN 1 na na na <0.001 and ALT >1.5*ULN, 0.852 0.67 1.084 0.192 level <=3*ULN >3*ULN 1.83 1.233 2.716 0.003

~~Table 4: Results of the Cox Regression Analysis of the Reduced Model for OS in Regorafeinb-Treated Patients (n=327).~~

Covariate Category HR LLCI ULCI P-value of P-value of estimate category covariate CAVD28 100 ng/mL 0.984 0.965 1.004 0.127 0.121 ECOG stage 0 1 na na Na 0.002 >=1 1.640 1.217 2.209 0.001 AFP level <400 ng/mL 1 na na Na 0 >=400 ng/mL 1.986 1.488 2.649 <0.001 Max of AST <= 1.5*ULN 1 na na Na 0.012 and ALT >1.5*ULN, 0.952 0.7 1.296 0.756 level <=3*ULN >3*ULN 2.084 1.270 3.422 0.004

Similarly, no definite E-R relationship was observed between CAVD28 and Time to progression (TTP). Univariate K-M curve didn’t show a trend of different survival across different exposure quartiles (Figure 19), nor did multivariate cox regression analysis identify a statistically significant relationship between CAVD28 and TTP after adjusting for significant baseline covariates (ECOG stage, Age category and AFP level) (Table 5).

Sensitivity analyses were conducted to assess the impact of excluded subjects due to missing data on the results. The sensitivity analysis resulted in similar parameter estimates, which indicated the analysis was robust with respect to the handling of missing data (i.e. exposure estimates, baseline covariates). This also held true for the selection of exposure metrics, as substituting CAVD28 with another exposure metric (CAVD1) did not result in different interpretations.

~~Reference ID: 4080727 Figure 19: (Univariate) Kaplan-Meier Curves for TTP Stratified by CAVD28 Exposure Quartiles and Placebo.~~

~~Table 5: Results of the Cox Regression Analysis of the Reduced Model for TTP in Regorafeinb-Treated Patients (n=315).~~

Covariate Category HR LLCI ULCI P-value of P-value of estimate category covariate CAVD28 100 ng/mL 0.995 0.978 1.013 0.600 0.599 ECOG stage 0 1 na na Na 0.002 >=1 1.577 1.198 2.075 0.001 AFP level <400 ng/mL 1 na na Na 0.018 >=400 ng/mL 1.366 1.057 1.765 0.017 Age <65 years 1 na na na 0.009 >=65 years 0.713 0.552 0.920 0.009

Reviewer’s comments: Extensive analyses using K-M and multivariate cox proportional hazard regression analysis were conducted by the applicant to characterize the E-R relationship for OS and TTP. Although there appears to be a trend of lower overall survival in the low exposure group, no statistically significant relationship

Reference ID: 4080727 was found between regorafenib exposure and OS or TTP after adjusting for baseline covariates. Thus, no evident and conclusive E-R relationship for efficacy was identified for regorafenib.

~~4.2.2 Exposure-Response for Safety~~

~~4.2.2.1 Overall treatment emergent AEs~~

The exposure-response for safety was first explored by evaluating the incidence rates of treatment- emergent adverse events (TEAEs) as a function of categories of estimated exposure metrics. Exposure metrics were derived from Pop-PK analysis and were described in section 4.4.1.

There was a trend of higher incidence rate of TEAEs ≥ Grade 3 during cycle 1 in the low regorafenib and free aggregate exposure groups (Figure 20). However, no statistically significant relationship between incidence rate of TEAEs ≥ Grade 3 and CAVD28 was identified in the logistic regression analysis (P-value = 0.074 and 0.292 for parent and free aggregate, respectively). It should also be noted that over the whole treatment period all of the patients experienced at least one TEAE. Similarly, no significant association was found between CAVD28 and incidence rate of ≥ grade 3 treatment-emergent serious adverse events (TESAEs) during cycle 1 by logistic regression (Figure 20) (P-value = 0.123 and 0.113 for parent and free aggregate, respectively).

~~Figure 20: Incidence Rate of TEAEs (Left)/TESAEs (Right) ≥ Grade 3 during Cycle 1 across Categories of Estimated Exposure of Regorafenib during Cycle 1~~

~~Reference ID: 4080727~~

~~(A) TEAEs (B) TESAEs~~

~~4.2.2 TEAEs of special interest~~

Exposure-response relationships were further investigated for TEAEs of special interest, namely hypertension, diarrhea, HFSR and asthenia/fatigue. There was no consistent exposure-dependent change in the overall incidence of any particular TEAE with increasing free aggregate exposure (Figure 21).

~~Figure 21: Incidence Rate of TEAEs of Special Interest (Hypertension, Diarrhea, HFSR and Fatigue) during Cycle 1 across Categories of Estimated Average Exposure of Free Aggregate in Cycle 1~~

~~(A) Hypertension (B) Diarrhea~~

~~Reference ID: 4080727~~

~~(C) HFSR (D) Fatigue~~

~~4.2.3 Laboratory data~~

The incidence rates of laboratory abnormalities of ALT, AST, platelet and total bilirubin (Grade ≥ 1) were compared across different categories of estimated average exposure of free aggregate in cycle 1 (Figure 22). There was no clear exposure-dependent increase in overall number of subjects with elevated AST or ALT with increasing free aggregate exposure of regorafenib. An increased incidence rate with number of subjects with elevated bilirubin or decreased platelet count was observed in higher exposure groups. Exploratory univariate logistic regression also identified a statically significant relationship between free aggregate exposure and occurrence of abnormal bilirubin (P-value=0.021) and platelet count (P- value=0.042). However, no exposure-dependent trend was observed in ≥ Grade 3 elevated bilirubin or decreased platelet count due to low number of observed events.

Reference ID: 4080727 Figure 22: Incidence Rate of Laboratory Abnormalities (Grade ≥ 1) of ALT, AST, Platelet and Total Bilirubin during Cycle 1 across Categories of Estimated Average Exposure of Free Aggregate in Cycle 1

~~(A) AST (B) ALT~~

~~(C) Bilirubin (B) Platelet count~~

Reviewer’s comments: Exposure-safety evaluations were explored with relevant safety data such as the incidence rate of overall TEAEs, the incidence rates of TEAE of special interest (Hypertension, Diarrhea, Fatigue and HFSR) and laboratory abnormalities (AST, ALT, bilirubin and platelet). No consistent and conclusive exposure- dependent relationships were identified between regorafenib and free aggregate exposures and the abovementioned safety data.

Reference ID: 4080727 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------YOUWEI N BI 04/06/2017

~~VADRYN PIERRE 04/06/2017~~

~~JIANG LIU 04/06/2017~~

~~JEANNE FOURIE ZIRKELBACH 04/06/2017~~

~~Reference ID: 4080727 CENTER FOR DRUG EVALUATION AND RESEARCH~~

~~APPLICATION NUMBER:~~

~~203085Orig1s007~~

~~OTHER REVIEW(S) MEMORANDUM~~

Date: April 17, 2017 From: Whitney S. Helms, PhD Pharmacology Supervisor Division of Hematology Oncology Toxicology for Division of Oncology Products 2 To: File for NDA # 203085 STIVARGA (regorafenib) Re: Labeling Supplement for Regorafenib (Supplment 7)

The Applicant provided references to support the inclusion of additional data in Section 12.1 (Mechanism of Action) of the label for regorafenib. Data to support the inclusion of CSFR1 as a clinically relevant target of regorafenib was previously reviewed during the original review of NDA 203085. In

Study A58227 the Applicant showed that regorafenib was able to bind CSF1R with a KD of 10 nM, within the range of binding for other kinases included in the label and a concentration that is clinically achievable at the recommended dose of 160 mg orally once daily for 21 days of a 28-day cycle. The originally reviewed xenograft data also specifically included studies using GIST and hepatocellular carcinoma tumors. Since these are the approved indications for regorafenib, the Applicant proposed naming these models in the label and FDA agreed.

The Applicant also submitted data from the literature describing the presence of tumor associated macrophages in syngeneic colorectal tumors orthotopically implanted in CD-1 mice. Beginning 4 days after orthotopic tumor implantation of a colorectal tumor model (CT26) in CD1-nude mice, animals were treated with regorafenib at a dose of 30 mg/kg for 10 days. The number of tumor associated macrophages, measured by staining with F4/80, decreased in regorafenib-treated mice compared to vehicle control or a VEGFR-2 antibody, DC101 (

~~Figure 1 A-B) 1 . More specifically, treatment with regorafenib resulted in lower levels of TIE2+ macrophages compared to the anti-VEGFR2 antibody (~~

~~Figure 1 C-D). The authors cite data indicating a role for TIE2-expressing macrophages in promoting angiogenesis and tumor progression2, 3.~~

1 Abou-Elkacem L, Arns S, Brix G, Gremse F, Zopf D, Kiessling F, Lederle W. Regorafenib inhibits growth, angiogenesis, and metastasis in a highly aggressive, orthotopic colon cancer model. Mol Cancer Ther. 2013 Jul;12(7):1322-311

~~2 Condeelis J, Pollard JW. Macrophages: obligate partners for tumor cell migration, invasion, and metastasis. Cell. 2006 Jan 27;124(2):263-6.~~

~~3 De Palma M, Naldini L. Angiopoietin-2 TIEs up macrophages in tumor angiogenesis. Clin Cancer Res 2011;17:5226–32~~

~~Reference ID: 4089862 Figure 1: Treatment with Regorafenib Reduces Macrophage, Including TIE2+ Macrophages, in Orthotopically Implanted Tumors~~

~~(Excerpted from Abou-Elkacem et. al)~~

This paper, thus, (b) (4) helps support the addition of a contribution of regorafenib to tumor immunity or maintenance of the tumor microenvironment, as stated in the originally approved label. The statement proposed by the Applicant, however, (b) (4)

The same study by Abou-Elkacem confirmed a decrease in metastatic lesions in mice treated with regorafenib compared to DC101 or vehicle control, similar to previous findings showing a decrease in metastatic activity in response to this drug. A second study using an orthotopic implantation model of tumor cells with mesenchymal stem cells also demonstrated a decrease in metastatic lesions in mice treated with regorafenib compared to control- treated animals4.

~~Figure 2: Inhibition of Metastases by Regorafenib~~

~~(Left: Excerpted from Abou-Elkacem et. al. Right: Excerpted from Takigawa et. al.)~~

4 Takigawa H, Kitadai Y, Shinagawa K, Yuge R, Higashi Y, Tanaka S, Yasui W, Chayama K. Multikinase inhibitor regorafenib inhibits the growth and metastasis of colon cancer with abundant stroma. Cancer Sci. 2016 May;107(5):601-8

~~Reference ID: 4089862 Current Proposed FDA Recommends Justification~~

Regorafenib is a small molecule Regorafenib is a small molecule Regorafenib is a small molecule inhibitor of See Disscussion above. inhibitor of multiple membrane- inhibitor of multiple membrane-bound multiple membrane-bound and intracellular The Applicant submitted bound and intracellular kinases and intracellular kinases involved in kinases involved in normal cellular functions additional literature involved in normal cellular functions normal cellular functions and in and in pathologic processes such as (b) (4) and in pathologic processes such as pathologic processes such as oncogenesis, tumor angiogenesis, oncogenesis, tumor angiogenesis, oncogenesis, tumor angiogenesis, metastasis and tumor immunity. In in vitro and maintenance of the tumor metastasis and tumor immunity. In in biochemical or cellular assays, regorafenib microenvironment. In in vitro vitro biochemical or cellular assays, or its major human active metabolites M-2 biochemical or cellular assays, regorafenib or its major human active and M-5 inhibited the activity of RET, regorafenib or its major human metabolites M-2 and M-5 inhibited the VEGFR1, VEGFR2, VEGFR3, KIT, PDGFR- active metabolites M-2 and M-5 activity of RET, VEGFR1, VEGFR2, alpha, PDGFR-beta, FGFR1, FGFR2, TIE2, inhibited the activity of RET, VEGFR1, VEGFR3, KIT, PDGFR-alpha, PDGFR-beta, DDR2, TrkA, Eph2A, RAF-1, BRAF, BRAF VEGFR2, VEGFR3, KIT, PDGFR-alpha, FGFR1, FGFR2, TIE2, DDR2, TrkA, Eph2A, V600E, SAPK2, PTK5, Abl and CSF1R) at PDGFR-beta, FGFR1, FGFR2, TIE2, RAF-1, BRAF, BRAF V600E, SAPK2, PTK5, concentrations of regorafenib that have DDR2, TrkA, Eph2A, RAF-1, BRAF, Abl and CSF1R) at concentrations of been achieved clinically. In in vivo models, BRAF V600E, SAPK2, PTK5, and Abl at regorafenib that have been achieved regorafenib demonstrated anti-angiogenic concentrations of regorafenib that clinically. In in vivo models, regorafenib activity in a rat tumor model and inhibition have been achieved clinically. In in demonstrated anti-angiogenic activity in of tumor growth in several mouse xenograft vivo models, regorafenib a rat tumor model, inhibition of tumor models including some for human colorectal demonstrated anti-angiogenic growth in several mouse xenograft carcinoma, gastrointestinal stromal and activity in a rat tumor model, and models including some for human hepatocellular carcinoma. Regorafenib also inhibition of tumor growth as well as colorectal carcinoma, gastrointestinal demonstrated antimetstatic activity in a anti-metastatic activity in several stromal and hepatocellular carcinoma, mouse xenograft model and two mouse mouse xenograft models including (b) (4) orthotopic models of human colorectal some for human colorectal carcinoma. carcinoma.

Reference ID: 4089862 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------WHITNEY S HELMS 04/26/2017

~~Reference ID: 4089862 FOOD AND DRUG ADMINISTRATION Center for Drug Evaluation and Research Office of Prescription Drug Promotion~~

~~****Pre-decisional Agency Information****~~

~~Memorandum~~

~~Date: April 26, 2017~~

~~To: Anuja Patel, MPH Senior Regulatory Health Project Manager Division of Oncology Products 2 Office of Hematology & Oncology Products~~

~~From: Carole C. Broadnax, R.Ph., Pharm.D. Regulatory Review Officer Office of Prescription Drug Promotion (OPDP)~~

~~Subject: Stivarga (regorafenib) tablets, for oral use NDA 203085 – 007 Addendum to OPDP Comments on proposed labeling~~

In response to the Division of Oncology Products 2’s (DOP 2) December 12, 2016, consult request, OPDP provided initial comments on April 18, 2017, for proposed labeling (package insert (PI) and patient package insert (PPI)) for Stivarga (regorafenib) tablets, for oral use.

This addendum is for OPDP’s additional comment (below) to DOP 2 for the Highlights section and for additional edits to Section 5.9 Wound Healing Complications of the proposed PI (attached) to further revise “regorafenib” to “STIVARGA” per the agreement reached during the April 24, 2017, tcon with Bayer. The additional Highlights comment and Section 5.9 edits were conveyed to DOP 2 via electronic mail and SharePoint on April 25, 2017.

OPDP notes that Highlights is inconsistent with Section 2.1 Dosage and Administration and that Section 12.3 Pharmacokinetics, Absorption (last paragraph, last sentence) and Section 14.1 Colorectal Cancer (second paragraph, second sentence) are inconsistent with Section 2.1. Should the sentences in Highlights, 12.3, and 14.1 be revised to be consistent with 2.1? For example, Highlights states, “Take STIVARGA with a low-fat meal. (2.1, 12.3);” however, Section 2.1 states, “Swallow tablet whole with water after a low-fat meal that contains less than 600 calories and less than 30% fat [see Clinical Pharmacology (12.3)].

The version of the proposed PI used in this review was sent to OPDP (Carole Broadnax) from DOP-2 (Anuja Patel) via a link to Sharepoint contained in electronic mail dated April 24, 2017, and is titled, “stivarga-draft-uspi-ccds9-hcc-tracked-word-version.docx.”

1 Reference ID: 4089377 Thank you for the opportunity to provide comments on these proposed materials. If you have any questions regarding this consult review, please contact Carole Broadnax at 301-796-0575 or [email protected].

~~28 Page(s) of Draft Labeling has been Withheld in Full as b4 (CCI/TS) immediately following this page~~

2 Reference ID: 4089377 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------CAROLE C BROADNAX 04/26/2017

Reference ID: 4089377 Department of Health and Human Services Public Health Service Food and Drug Administration Center for Drug Evaluation and Research Office of Medical Policy Initiatives Division of Medical Policy Programs

~~PATIENT LABELING REVIEW~~

~~Date: April 20, 2017~~

~~To: Patricia Keegan, MD Director Division of Oncology Products 2 (DOP2)~~

~~Through: LaShawn Griffiths, MSHS-PH, BSN, RN Associate Director for Patient Labeling Division of Medical Policy Programs (DMPP)~~

~~Barbara Fuller, RN, MSN, CWOCN Team Leader, Patient Labeling Division of Medical Policy Programs (DMPP)~~

~~From: Morgan Walker, PharmD, MBA, CPH Patient Labeling Reviewer Division of Medical Policy Programs (DMPP) Subject: Focused Review of Patient Labeling: Patient Package Insert (PPI)~~

~~Drug Name (established STIVARGA (regorafenib) name): Dosage Form and Route: tablets, for oral use Application NDA 203085/S-007 Type/Number/Supplement: Applicant: Bayer HealthCare Pharmaceuticals Inc.~~

~~Reference ID: 4086969 1 INTRODUCTION~~

On October 31, 2016, Bayer HealthCare Pharmaceuticals Inc. submitted for the Agency’s review a Prior Approval Supplement to New Drug Application (NDA) 203085/S-007 for STIVARGA (regorafenib) tablets. This labeling supplement provides a proposed new indication for the treatment of patients with hepatocellular (b) (4) carcinoma (HCC) who have been previously treated with . STIVARGA (regorafenib) tablets was originally approved on September 27, 2012 with the indication for the treatment of patients with metastatic colorectal cancer (CRC) who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan based chemotherapy, an anti-VEGF therapy, and, if KRAS wild type, an anti-EGFR therapy. On February 25, 2013 STIVARGA (regorafenib) tablets was approved under NDA 204369 for the treatment of patients with locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate. This focused review is written by the Division of Medical Policy Programs (DMPP) in response to a request by the Division of Oncology Products 2 (DOP2) on December 12, 2016 for DMPP to provide a focused review of the Applicant’s proposed Patient Package Insert (PPI) for STIVARGA (regorafenib) tablets.

2 MATERIAL REVIEWED • Draft STIVARGA (regorafenib) tablets PPI received on October 31, 2016, and received by DMPP on April 10, 2017. • Draft STIVARGA (regorafenib) tablets Prescribing Information (PI) received on October 31, 2016, revised by the Review Division throughout the review cycle, and received by DMPP on April 10, 2017.

~~3 REVIEW METHODS In our focused review of the PPI we: • simplified wording and clarified concepts where possible • ensured that the PPI is consistent with the Prescribing Information (PI)~~

~~4 CONCLUSIONS The PPI is acceptable with our recommended changes.~~

5 RECOMMENDATIONS • Please send these comments to the Applicant and copy DMPP on the correspondence. • Our focused review of the PPI is appended to this memorandum. Consult DMPP regarding any additional revisions made to the PI to determine if corresponding revisions need to be made to the PPI. Please let us know if you have any questions.

~~6 Page(s) of Draft Labeling has been Withheld in Full as b4 (CCI/TS) immediately following this page~~

Reference ID: 4086969 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------MORGAN A WALKER 04/20/2017

~~BARBARA A FULLER 04/20/2017~~

~~LASHAWN M GRIFFITHS 04/20/2017~~

~~Reference ID: 4086969 FOOD AND DRUG ADMINISTRATION Center for Drug Evaluation and Research Office of Prescription Drug Promotion~~

~~****Pre-decisional Agency Information****~~

~~Memorandum~~

~~Date: April 18, 2017~~

~~To: Anuja Patel, MPH Senior Regulatory Health Project Manager Division of Oncology Products 2 Office of Hematology & Oncology Products~~

~~From: Carole C. Broadnax, R.Ph., Pharm.D. Regulatory Review Officer Office of Prescription Drug Promotion (OPDP)~~

~~Subject: Stivarga (regorafenib) tablets, for oral use NDA 203085 – 007 OPDP Comments on proposed labeling~~

In response to the Division of Oncology Products 2’s (DOP 2) December 12, 2016, consult request, OPDP has reviewed proposed labeling (package insert (PI) and patient package insert (PPI)) for Stivarga (regorafenib) tablets, for oral use.

This supplemental application proposes to revise the PI to: • include a new indication for the treatment of patients with hepatocellular carcinoma (b) (4) (HCC) who have been previously treated with ; • update Section 5 Warnings and Precautions, and Section 6 Adverse Reactions, with pooled data from four phase 3 trials (CORRECT (#14387), GRID (#14874), RESORCE (#15982), and CONCUR (#15808) following completion of RESORCE trial (#15982); • add a new subsection, 5.2 Infections, under Warnings and Precautions based on data from the pooled phase 3 trials; • update subsection 12.1 Mechanism of Action, to include language on metastasis and tumor immunity; and, • revise the PPI for consistency with revisions proposed in the PI.

Comments on the proposed labeling are based on the substantially complete labeling dated April 10, 2017, entitled “sNDA 203085 S 007 FDA Preliminary edits to PI received 6 Jan 17.docx.” OPDP’s comments on the proposed labeling are provided directly on the marked version below.

~~If you have any questions regarding this consult review, please contact Carole Broadnax at 301-796-0575 or [email protected].~~

~~Thank you for the opportunity to provide comments on these proposed materials. 27 Page(s) of Draft Labeling has been Withheld in Full as b4 (CCI/TS) immediately following this page~~

1 Reference ID: 4085734 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------CAROLE C BROADNAX 04/18/2017

Reference ID: 4085734 LABELING REVIEW Division of Medication Error Prevention and Analysis (DMEPA) Office of Medication Error Prevention and Risk Management (OMEPRM) Office of Surveillance and Epidemiology (OSE) Center for Drug Evaluation and Research (CDER)

~~*** This document contains proprietary information that cannot be released to the public***~~

Date of This Review: December 30, 2016 Requesting Office or Division: Division of Oncology Products 2 (DOP2) Application Type and Number: NDA 203085/S-007 Product Name and Strength: Stivarga (regorafenib) Tablets, 40 mg Product Type: Single Ingredient Product Rx or OTC: Rx Applicant/Sponsor Name: Bayer Health Care Pharmaceuticals Inc. (Bayer) Submission Date: October 31, 2016 OSE RCM #: 2016-2550 DMEPA Primary Reviewer: Otto L. Townsend, PharmD DMEPA Team Leader: Chi-Ming (Alice) Tu, PharmD, BCPS

Reference ID: 4035420 1 REASON FOR REVIEW This review responds to a request from the Division of Oncology Products 2 (DOP2) to evaluate Section 2 of the proposed Prescribing Information (PI) submitted by Bayer for areas of vulnerability that could lead to medication errors. The PI was submitted as part of an efficacy supplement which, if approved, would expand the indication of Stivarga to include the treatment of patients with hepatocellular carcinoma who have been previously treated with (b) (4) .

2 MATERIALS REVIEWED We considered the materials listed in Table 1 for this review. The Appendices provide the methods and results for each material reviewed. Table 1. Materials Considered for this Label and Labeling Review Material Reviewed Appendix Section (for Methods and Results) Product Information/Prescribing Information A Previous DMEPA Reviews B Human Factors Study C – N/A ISMP Newsletters D FDA Adverse Event Reporting System (FAERS)* E Other F – N/A Labels and Labeling G N/A=not applicable for this review *We do not typically search FAERS for label and labeling reviews unless we are aware of medication errors through our routine postmarket safety surveillance

3 OVERALL ASSESSMENT OF THE MATERIALS REVIEWED If approved, this efficacy supplement (S-007) will expand the indication of Stivarga (regorafenib) to include the treatment of patients with hepatocellular cancer using the same dosing recommendations as the currently approved indications. The submission included proposed changes to the prescribing information, but did not include any proposed changes to the container label or carton labeling. Bayer proposed to add the following introductory statements in Section 2.2 (Dose Modification) of the PI: (b) (4)

Reference ID: 4035420 Bayer provided the following rationale for the addition of these introductory statements: Rationale (b) (4) Sponsor believes it is appropriate to provide the prescriber with a general overview of the dose modification strategy for regorafenib, including mention of the dose modifications steps (b) (4) prior to offering detailed guidance regarding the management of specific adverse events. Such an approach is expected to improve the readability and therefore the understanding of the dose modification guidance. Rationale (b) (4)

~~(b) (4)~~

As part of our risk assessment, we also considered our recommendations and surveillance plan from our recent Postmarket Reviews that evaluated the potential overdose risk that is present when patients are prescribed 120 mg or 80 mg as dose reductions from the standard dose of 160 mg (see Appendix B for a summary or the referenced reviews for more details). Our gap search of FAERS and ISMP Newsletters did not identify any overdose medication errors associated with the availability of extra tablets when prescribed the 120 mg or 80 mg dose. We will continue to monitor medication error reports related to Stivarga.

~~4 CONCLUSION & RECOMMENDATIONS The Applicant suggested (b) (4) . We provide recommendations for consideration in Section 4.1 below.~~

4.1 RECOMMENDATIONS FOR THE DIVISION A. Prescribing Information 1. We defer the decision on the appropriateness of (b) (4) to the review team. If the review team determines that (b) (4) is appropriate, we recommend that the Applicant propose specific dosing guidelines in the Dosing and Administration section of the PI. For example, a new subsection (b) (4)

~~(b) (4)~~

~~Reference ID: 4035420 APPEARS THIS WAY ON ORIGINAL~~

~~Reference ID: 4035420 APPENDICES: METHODS & RESULTS FOR EACH MATERIALS REVIEWED~~

APPENDIX A. PRODUCT INFORMATION/PRESCRIBING INFORMATION Table 2 presents relevant product information for Stivarga that Bayer submitted on October 31, 2016. Table 2. Relevant Product Information for Stivarga Initial Approval Date September 27, 2012 Active Ingredient Regorafenib Indication Current: Colorectal Cancer and Gastrointestinal Stromal Tumors

~~Proposed: treatment of patients with hepatocellular carcinoma who have been previously treated with (b) (4)~~

Route of Administration Oral Dosage Form Tablets Strength 40 mg Dose and Frequency 160 mg (4 x 40 mg tablets) by mouth once daily for the first 21 days of each 28-day cycle. How Supplied Carton containing three bottles, with each bottle containing 28 tablets, for a total of 84 tablets per carton. Storage Store at 25°C (77°F); excursions are permitted from 15 to 30°C (59 to 86°F) [see USP Controlled Room Temperature].

~~Store tablets in the original bottle and do not remove the desiccant. Keep the bottle tightly closed after first opening.~~

Discard any unused tablets 7 weeks after opening the bottle. Dispose of unused tablets in accordance with local requirements. Container Closure Plastic white opaque bottle with child-resistant closure.

Reference ID: 4035420 APPENDIX B. PREVIOUS DMEPA REVIEWS B.1 Methods On December 6, 2016, we searched the L:drive and AIMS using the terms, “Stivarga” and “regorafenib” to identify reviews previously performed by DMEPA.

B.2 Results Our search identified 3 previous reviewsb,c,d. The 2 most recent reviews were postmarket reviews that addressed a safety issue involving the potential overdose risk that is present when patients are prescribed 120 mg or 80 mg as dose reductions from the standard dose of 160 mg. A General Advise letter from DOP2 was issued to the Applicant requesting the following: 1. A summary of any medication errors or product complaints that Bayer has received that are associated with the package size of Stivarga. 2. A rationale for only supplying Stivarga in a 28-count bottle, which provides more tablets than the quantity required for dose modification. 3. Consideration for changing the package size if Stivarga to a 21- count bottle (carton containing 4 bottles). In response to the General Advice letter from DOP2, Bayer determined there was not a need to change from a 28-count to 21-count bottle because they were not able to identify medication errors or product complaints that could be attributed to the availability of additional tablets when a patient is taking either a reduced dose of 120 mg or 80 mg. They proposed to continue monitoring this issue. In the absence of any medication errors related to package size, we found Bayer’s proposal acceptable and we continue to monitor for medication errors related to Stivarga.

b Schlick, J. Label, Labeling, and Packaging Review for Stivarga (NDA 203085). Silver Spring (MD): FDA, CDER, OSE, DMEPA (US); 2012 JUL 25. RCM No.: 2012-1082. c Townsend, O. Postmarket Medication Error Memorandum for Stivarga (NDA 203085). Silver Spring (MD): FDA, CDER, OSE, DMEPA (US); 2015 OCT 07. RCM No.: 2015-1809. d Stewart, J. Postmarket Medication Error Memorandum for Stivarga (NDA 203085). Silver Spring (MD): FDA, CDER, OSE, DMEPA (US); 2016 MAY 24. RCM No.: 2016-998.

Reference ID: 4035420 APPENDIX D. ISMP NEWSLETTERS D.1 Methods On December 6, 2016, we searched the Institute for Safe Medication Practices (ISMP) newsletters using the criteria below, and then individually reviewed each newsletter. We limited our analysis to newsletters that described medication errors or actions possibly associated with the label and labeling. ISMP Newsletters Search Strategy ISMP Newsletter(s) Acute Care, Community and or Nursing Search Strategy and Terms Match Any of the Words: Stivarga, regorafenib

D.2 Results Our search of the Institute for Safe Medication Practices (ISMP) newsletters did not yield any newsletters that described medication errors or actions possibly associated with the label and labeling of Stivarga.

Reference ID: 4035420 APPENDIX E. FDA ADVERSE EVENT REPORTING SYSTEM (FAERS) E.1 Methods We performed a gap search of the FDA Adverse Event Reporting System (FAERS) on December 6, 2016 using the criteria in Table 3, and then individually reviewed each case. We limited our analysis to cases that described errors possibly associated with the label and labeling. We used the NCC MERP Taxonomy of Medication Errors to code the type and factors contributing to the errors when sufficient information was provided by the reporter.e Table 3: FAERS Search Strategy Initial FDA Receive Dates May 1, 2016 to November 30, 2016 Product Name Stivarga Event (MedDRA Terms) Medication errors SMQ (narrow)

E.2 Results Our search retrieved 10 cases, but after further evaluation, we didn’t identify any medication error cases that were relevant for this review and could be addressed by labels and labeling revisions. E.3 Description of FAERS The FDA Adverse Event Reporting System (FAERS) is a database that contains information on adverse event and medication error reports submitted to FDA. The database is designed to support the FDA's postmarket safety surveillance program for drug and therapeutic biologic products. The informatic structure of the FAERS database adheres to the international safety reporting guidance issued by the International Conference on Harmonisation. FDA’s Office of Surveillance and Epidemiology codes adverse events and medication errors to terms in the Medical Dictionary for Regulatory Activities (MedDRA) terminology. Product names are coded using the FAERS Product Dictionary. More information about FAERS can be found at: http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Surveillance/AdverseD rugEffects/default.htm.

~~e The National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) Taxonomy of Medication Errors. Website http://www.nccmerp.org/pdf/taxo2001-07-31.pdf.~~

Reference ID: 4035420 APPENDIX G. LABELS AND LABELING G.1 List of Labels and Labeling Reviewed Using the principles of human factors and Failure Mode and Effects Analysis,f along with postmarket medication error data, we reviewed the following Stivarga labels and labeling submitted by Bayer on October 31, 2016.

~~ Prescribing Information~~

~~f Institute for Healthcare Improvement (IHI). Failure Modes and Effects Analysis. Boston. IHI:2004.~~

Reference ID: 4035420 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------OTTO L TOWNSEND 12/30/2016

~~CHI-MING TU 12/30/2016~~

~~Reference ID: 4035420 CENTER FOR DRUG EVALUATION AND RESEARCH~~

~~APPLICATION NUMBER:~~

~~203085Orig1s007~~

~~ADMINISTRATIVE and CORRESPONDENCE DOCUMENTS~~

~~DEPARTMENT OF HEALTH AND HUMAN SERVICES~~

~~Food and Drug Administration Silver Spring MD 20993~~

~~IND 075642 MEETING MINUTES~~

~~Bayer HealthCare Pharmaceuticals, Inc. Attention: Lisa Chao, Ph.D. Deputy Director, Regulatory Affairs 100 Bayer Boulevard, P.O. Box 915 Whippany, NJ 07981-0915~~

~~Dear Dr. Chao:~~

~~Please refer to your Investigational New Drug Application (IND) submitted under section 505(i) of the Federal Food, Drug, and Cosmetic Act for regorafenib.~~

We also refer to the meeting between representatives of your firm and the FDA on July 21, 2016. The purpose of the meeting was to discuss the efficacy and safety results of Study 15982, entitled “A Randomized, Double-Blind, Placebo-Controlled, Multicenter Phase III Study of Regorafenib in Patients with Hepatocellular Carcinoma (HCC) After Sorafenib”; and to seek agreement on the contents of the sNDA and potential rolling review plan.

~~A copy of the official minutes of the meeting is enclosed for your information. Please notify us of any significant differences in understanding regarding the meeting outcomes.~~

~~If you have any questions, call me at (301) 796-3074.~~

~~Sincerely,~~

~~{See appended electronic signature page}~~

~~Idara Udoh, M.S. Senior Regulatory Health Project Manager Division of Oncology Products 2 Office of Hematology and Oncology Products Center for Drug Evaluation and Research~~

~~Enclosure: Meeting Minutes~~

~~Reference ID: 3964863 FOOD AND DRUG ADMINISTRATION CENTER FOR DRUG EVALUATION AND RESEARCH~~

~~MEMORANDUM OF MEETING MINUTES~~

~~Meeting Type: Type B Meeting Category: Pre-sNDA~~

~~Meeting Date and Time: July 21, 2016; 1:00 PM – 2:00 PM, EST Meeting Location: White Oak Building 22, Conference Room: 1421~~

Application Number: IND 075642 Product Name: Regorafenib Indication: Treatment of patients with hepatocellular carcinoma (HCC) who (b) (4) have been previously treated with Sponsor/Applicant Name: Bayer HealthCare Pharmaceuticals, Inc.

~~Meeting Chair: Steven Lemery Meeting Recorder: Idara Udoh~~

FDA ATTENDEES Center for Drug Evaluation and Research, Division of Oncology Products 2 (DOP 2) Martha Donoghue, Acting Associate Director Steven Lemery, Clinical Team Leader Damiette Smit, Clinical Reviewer Lola Fashoyin-Aje, Clinical Reviewer Monica Hughes, Chief, Project Management Staff Idara Udoh, Senior Regulatory Health Project Manager

Center for Drug Evaluation and Research, Office of Clinical Pharmacology Hong Zhao, Clinical Pharmacology Team Leader Jeanne Fourie Zirkelbach, Clinical Pharmacology Team Leader Jingyu Yu, Clinical Pharmacology Reviewer Vadryn Pierre, Clinical Pharmacology Reviewer

~~Office of Pharmaceutical Quality (OPQ), Office of Lifecycle Drug Products Zedong Dong, Quality Assessment Lead (Acting)~~

~~SPONSOR ATTENDEES Bayer HealthCare Pharmaceuticals, Inc. Gerold Meinhardt, Clinical Development Mark Rutstein, Clinical Development Adriaan Cleton, Clinical Pharmacology~~

~~Reference ID: 3964863 IND 075642 Page 2~~

Zuzana Jirakova Trnkova, Clinical Pharmacology Christian Kappler, Clinical Statistics Hui-Talia Zhang, Pharmacovigilance Kathleen Schostack, Program Head Lisa Chao, Regulatory Affairs Birgit Wolf, Regulatory Affairs Philip Johnson, Regulatory Affairs Gerhard Schlueter, Regulatory Affairs Marie-Aude Le Berre

~~MEETING PURPOSE~~

On May 20, 2016, Bayer HealthCare Pharmaceuticals, Inc. (“Bayer”) requested a Type B pre- supplemental New Drug Application (sNDA) meeting to discuss the efficacy and safety results of Study 15982 (RESORCE), entitled “A Randomized, Double-Blind, Placebo-Controlled, Multicenter Phase III Study of Regorafenib in Patients with Hepatocellular Carcinoma (HCC) After Sorafenib” and to seek agreement on the contents of an sNDA and potential rolling review plan. Bayer plans to submit the sNDA in August/September 2016.

~~The meeting package was received on June 20, 2016. FDA sent Preliminary Comments to Bayer on July 18, 2016.~~

~~PROPOSED INDICATION~~

~~For the treatment of patients with hepatocellular carcinoma (HCC) who have been previously (b) (4) treated with .~~

~~BACKGROUND~~

~~Regulatory History~~

Stivarga (regorafenib) is approved for the following indications: • the treatment of patients with metastatic colorectal cancer (mCRC) who have been previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti- VEGF therapy, and, if KRAS wild type, an anti-EGFR therapy • the treatment of patients with locally advanced, unresectable or metastatic gastrointestinal stromal tumor (GIST) who have been previously treated with imatinib mesylate and sunitinib malate.

On December 23, 2015, Bayer submitted a meeting request to obtain FDA feedback on the format and content of a planned sNDA based on the results of Study 15982, entitled “A Randomized, Double Blind, Placebo-Controlled, Multicenter Phase III Study of Regorafenib in Patients with Hepatocellular Carcinoma (HCC) after Sorafenib.” A final written response (WRO) was issued on March 4, 2016.

~~Reference ID: 3964863 IND 075642 Page3~~

~~~.4 {l>H 4) On June 15, 2016, FDA received Cb>< request Fast Track CbTC4l for treatment of atients with h~atocellular carcinoma (HCC) who have been ,_p-re-v.. 10-- u-sly treated with Cb)l~~

~~Clinical~~

Bayer has conducted the following studies to suppo1i a marketing application for regorafenib in HCC: • Study 11651 (an open-label, non-randomized, dose-escalating study) • Study 14596 (a single-arm, multicenter, open-label safety study) and • Study 15982 (a randomized, double blind, placebo-controlled, multicenter study).

Study 15982 will serve as the single pivotal clinical trial suppo1i ing the safety and efficacy evaluation of the proposed sNDA for the treatment of patients with HCC who have previously received sorafenib.

Study 14596 is a single-aim, multicenter safety study of regorafenib in patients with HCC (Child-Pugh A) after treatment with sorafenib. Thniy-six patients were enrolled. All patients experienced at least one treatment-emergent adverse event (TEAE). Three patients (8.3%) experienced a serious adverse event (SAE) dete1mined to be drng-related (fatigue, diaiThea and hematoma). Five patients died with one cause of death detennined to be drug-related (hematoma). The most frequent TEAE's included fatigue (69.4%), diaiThea (52.5%) and hand foot skin reaction (HFSR; 60%). The most frequent TEAE's of Grade 3 or greater severity included fatigue (22.2%), HFSR (13.9%) and hyperbilirnbineinia (11.1 %). Median overall survival (OS) was 13.8 months (range of 1.4 - 28.9 months) and time to progression (TTP) was 4.3 months. Thniy-one patients were evaluable for response. One patient had a paliial response and 25 patients had stable disease.

Study 15982 is a randomized, double-blind, multicenter, placebo-controlled trial that enrolled patients with HCC after they previously received sorafenib. A total of 560 patients were planned to be randoinized. Randomization was stratified by region (Asian vs. rest of world), ECOG PS (0 vs. 1), AFP level (<400 ng/mL vs. ~400 ng/mL), extrahepatic disease (presence vs. absence), and macrovascular invasion (presence vs. absence). A total of 573 patients were randomly assigned in a 2: 1 ratio to two treatment aims: • Experimental: regorafenib 160 mg orally, once daily, for 3 weeks of eve1y 4 week (28- day) cycle (i.e. 3 weeks on/1 week off) plus best suppo1i ive cai·e (BSC) • Control: identical placebo tablets following the same treatment schedule plus BSC.

~~Reference ID: 3964863 IND 075642 Page4~~

~~Treatment continued until disease progression, unacceptable treatment-related toxicity, or patient or physician decision to discontinue.~~

The primary endpoint was OS. Assuming that the median OS is 8 months in the control ann and 11.4 months in the experimental aim , a total of 370 events were needed to detect a hazard ratio (HR) of 0.70 with 90% power at a 1-sided alpha level of 2.5%. The primary analysis was a strntified log-rank test perfonned on the intent-to-treat population. One interim analysis of OS was planned after 111 (30%) events (for futility only) . The futility stopping boundary was specified as non-binding. Major secondaiy endpoints included progression-free smvival (PFS), TTP, and objective response rate (ORR). A hierai·chical procedure was proposed to adjust for multiplicity in testing the secondaiy endpoints with the order of PFS~ TTP.

The primary analysis of OS for Study 15982 was triggered after 372 events were obseived. The efficacy results ai·e: OS: HR 0.62 (95% CI: 0.5, 0.78), p<0.0001, median OS of 10.6 in the regorafenib aim and 7.8 months for placebo; PFS (mRECIST): HR 0.46 (95% CI: 0.37, 0.56), p<0.0001, median PFS 3. 1 and 1.5 months; ORR (mRECIST): 10.6% and 4.1%, p=0.0047.

All patients treated with regorafenib and nearly all patients treated with placebo (92.7%) had at least one TEAE. Of these, 10.4% vs. 2.6% were detennined to be both drng-related and serious in the regorafenib vs. placebo aims, respectively. Grade 5 events occmTed in 13.4% of patients treated with regorafenib and in 19.7% of patients treated with placebo. The most collllllon adverse events of Grade 3 or greater severity were HFSR (12.3% vs. 0.5% in the regorafenib vs. placebo group, respectively), dianhea (3 .2% vs. 0%), decreased appetite (2 .7% vs. 1.6%), hypeitension (1 4.7% vs. 4.7%), fatigue (5 .9% vs. 3.6%), AST increased (11 % vs. 11 .4%), hyperbilirnbineinia (7.5% vs. 9.3%), and abdominal pain (2 .7% vs. 2.6%).

~~DISCUSSION OF FDA RESPONSES TO SPONSOR QUESTIONS~~

~~Question 1: Study 15982 to Support sNDA Filing Bayer's position on Question #I provided in Section 9.1(page35) ofbriefing package.~~

1. Does the Division agree that the results from Study 15982, together with the results from the Phase 2 Study 14596, provide sufficient information for evaluation of the efficacy and safety of regorafenib in order to support an sNDA flling for the following indication: "Stivarga is indicated for the treatment of patients with he atocellular carcinoma (HCC) who have been previously treated with (b]{4 "?

FDA Response: FDA agrees that the results of both studies provide sufficient infoimation to allow FDA to evaluate the efficacy and safety of re orafenib. However, FDA does not agree wit. h t h e 12!2. ose d m. dicahon." . >H4lL----

~~Bayer's Response (Received via email on July 20, 2016): Bayer would like to discuss our rationale for the proposed indication. In that context, we would also like to discuss the~~

~~Reference ID: 3964863 IND 075642 Page 5~~

~~(b)(4l otential role of real-world evidence to suppo1t our proposed indication ------~~

Follow-up Question to FDA: Is our agreement to revise the indication as FDA has requested by email from Idara Udoh on July 12, 2016, a prerequisite for granting of Fast Track Designation? FDA proposal: "For the treatment ofE. atients with {l>H" ~ hepatocellular carcinoma (HCC) who have (b) 4 JJrBii'iel:tSI,' f.retlfe

~~Discussion During Meeting: FDA stated that the Fast Track designation if ggnted) would need to incorporate rior sorafenib treatment in the indication statement (b)14l~~

~~that Bayer could propose alternative labeling (with justification suppo1t ing their proposed approach) at the time of the original sNDA submission.~~

~~In regards to real-world evidence, FDA stated that Bayer could submit a proposal for review and feedback.~~

~~Question 2: Clinical Pharmacology Bayer's position on Question #2 provided in Section 9.2 (page 36) ofbriefing package.~~

~~2. Does the Division agree with Bayer's planned exposure-response and dose modification analyses?~~

FDA Response: In general, the planned exposure-response and dose modification analyses are acceptable. FDA recommends that Bayer conduct multivariate Cox regression and/or logistic regression analyses, and includes the analysis repo1t in the sNDA if the graphical exploration indicates any positive exposure-response relationship for efficacy.

Bayer's Response (Received via email on July 20, 2016): Analysis of the exposure response is currently ongoing. Bayer would like to get an understanding from FDA of their definition of "any positive exposure-response relationship for efficacy".

Discussion During Meeting: FDA clarified that an exposure-response effect for efficacy may exist if there is a clinically relevant, visual separation between Kaplan-Meier survival cmves in exposure groups for time-to-progression or overall smv ival. If such a separation is obse1ved, a multivariate analysis would be recommended. Bayer asked for clarification regarding baseline characteristics of interest. FDA confomed that the stratification factors should be considered for the explorato1y post hoc analysis; however, other impo1tant prognostic factors could also be identified and incorporated into the analysis.

~~Reference ID: 3964863 IND 075642 Page 6~~

~~Question 3: Applicability of Data to US Population Bayer’s position on Question #3 provided in Section 9.3 (page 39) of briefing package.~~

3. Does the Division concur with Bayer’s analyses for US versus non-US subjects (outlined in the US-specific SAP), as well as for subjects from Asian vs Non-Asian countries (outlined in the global SAP for Study 15982) and agree that these analyses are adequate to demonstrate applicability of this study to the US population?

~~FDA Response: The proposed method is acceptable; however, whether the results are adequate to support the applicability to the US population will be determined during the review of the sNDA.~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

~~Question 4: Clinical Benefit of Continued Treatment Past Progression Bayer’s position on Question #4 provided in Section 9.4 (page 41) of briefing package.~~

~~4. Does the Division agree with Bayer’s proposed approach to address the topic of treatment benefit post-progression?~~

FDA Response: Although we do not object to Bayer’s proposed approach, we do not agree that the approach is sufficient to demonstrate that regorafenib provided clinical benefit in patients who experienced radiographic progression (e.g., based on absence of deterioration of ECOG performance status).

Bayer’s Response (Received via email on July 20, 2016): Since Bayer will not be seeking label claims of clinical benefit beyond progression, we would like to request additional understanding of FDA’s feedback.

~~Discussion During Meeting: FDA and Bayer agree that the analyses based on this information would be considered exploratory.~~

~~Reference ID: 3964863 IND 075642 Page 7~~

~~Question 5: Case Report Forms (CRFs) and Narratives Bayer’s position on Question #5 provided in Section 9.5 (page 42) of briefing package.~~

~~5. Does the Division agree with our planned submission of these CRFs and narratives?~~

FDA Response: No. Please also provide narrative summaries for patients who discontinued receiving regorafenib due to loss of follow-up, physician decision, or subject decision (or for administrative or other reasons not related to disease progression). Please also be prepared to submit any additional CRFs or narratives upon request.

Bayer’s Response (Received via email on July 20, 2016): Bayer agrees to provide the additional narratives. Are these additional narrative summaries a new standard request for oncology applications?

Discussion During Meeting: FDA could not confirm that this is standard policy; however, additional CRFs or narratives are often requested as part of an NDA or sNDA submission to assist with the safety review.

~~Question 6: Rolling Review Bayer’s position on Question #6 provided in Section 9.6 (page 43) of briefing package.~~

~~6. Does the Division agree with this proposal of submitting the Module 2 and Module 5 clinical documents in a staggered manner in the context of the rolling submission?~~

(b) (4) FDA Response: If Bayer receives Fast Track based on the results of Study 15982, FDA agrees that Bayer can submit the clinical documents in a staggered manner in the context of a Rolling Review submission according to the following schedule proposed in the meeting package:

~~• August 2016: - Module 5: Study 15982 Final Study Report, Population PK Final Study Report, PK/PD Final Study Report, and Electronic Datasets. - Module 1 (except draft label)~~

• September 2016: - Module 5: Module 5.3.5.3 ISE and ISS, Electronic Datasets for Integrated Safety Analysis, Study 15982 US-Specific Report, and OSI requests. - Module 2: Module 2.7.2, 2.7.3, 2.7.4, and 2.6. - Module 1 (draft label).

(b) (4) If Bayer receives Fast Track , please submit a formal request for Rolling Review with the agreed upon schedule as an amendment to the IND (with attached Form FDA 1571). Clearly identify the submission as a REQUEST FOR SUBMISSION OF PORTIONS OF AN APPLICATION in bold, uppercase letters.

~~Reference ID: 3964863 IND 075642 Page 8~~

Bayer’s Response (Received via email on July 20, 2016): Bayer would like to inform FDA that we decided to reopen the database in order to address minor data issues that have no impact on the interpretation of the efficacy, safety and PK results, as well as on the benefit/risk profile of regorafenib. As a result, there will be a delay of approximately 1.5 months in the sNDA submission (for each wave of the rolling review) and the potential shifting of PK/PD and population PK reports into Wave 2. The revised rolling review strategy is as follows:

• Wave 1 (approx. end Sept/beginning Oct 2016): • Module 5: Study 15982 Final Study Report, Population PK Final Study Report, PK/PD Final Study Report, and Electronic Datasets for Study 15982 and Population PK. • Module 1 (except draft label) • Wave 2 (approx. end of Oct/beginning Nov 2016): • Module 5: Module 5.3.5.3 ISE and ISS, Electronic Datasets for Integrated Safety Analysis, Population PK Final Study Report, PK/PD Final Study Report, Electronic data sets for PK/PD, Study 15982 US-Specific Report, and OSI requests. • Module 2: Module 2.7.2, 2.7.3, 2.7.4, and 2.5. • Module 1 (draft label)

~~Does FDA agree with revised rolling review strategy?~~

Discussion During Meeting: FDA agrees that Bayer can submit the formal request for Rolling Review with the amended rolling submission strategy. FDA requested that Bayer document the data issues uncovered and how they were resolved.

~~Question 7: Applicant Orientation Meeting~~

~~7. Given that Bayer will present the key efficacy and safety data from Study 15982 at the Pre-sNDA meeting, does the Division anticipate requesting an Applicant Orientation meeting for the HCC sNDA?~~

~~FDA Response: No; however, FDA will contact Bayer if the Agency determines that the application would benefit from an Application Orientation meeting.~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

~~Reference ID: 3964863 IND 075642 Page 9~~

~~ADDITIONAL FDA COMMENTS~~

~~Clinical Pharmacology~~

~~In the sNDA submission:~~

8. Provide a justification for the exposure metrics that are used in the exposure-response analyses. Drug exposure to be used in the analyses may include but not be limited to trough concentration at steady-state, maximum concentration at steady-state, average concentration at steady-state or trough concentration after the first dose. The use of either the parent drug exposure or metabolites exposure, or both, should be justified.

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

9. Include the following in the exposure-response analysis reports: a. A summary of baseline characteristics including but not limited to demographics, disease features and lab measurements, for all patients included in the analysis and subgroups based on drug exposures.

Bayer’s Response (Received via email on July 20, 2016): Please see the following list of baseline characteristics that are planned to be incorporated in the analysis. This list can also be found in Section 6.1 of the Integrated Analysis SAP for the exposure- response analysis (Appendix 1 of the pre-sNDA Briefing Package). Bayer requests confirmation that this list is acceptable to FDA.

“Summary statistics (n, mean, SD, minimum, median, maximum) will be presented for the PKPD set for the variables age, weight, height, baseline body mass index (BMI), baseline ALT, baseline AST, baseline total bilirubin, and baseline platelets. If BMI is not available, it will be re-calculated if possible as baseline weight (kg) divided by height (m) squared. Frequency counts (n, %) will be presented for the variables sex, race group 1 (Asian vs Non-Asian), race group 2 (Japanese vs Chinese vs other Asian versus Non-Asian), ECOG performance status, baseline CTCAE grades for ALT, AST, total bilirubin, platelets, Child-Pugh score at baseline (A5 vs A6), hepatic function at baseline (maximum of baseline AST and baseline ALT ≤ 1.5*ULN vs > 1.5*ULN to ≤ 3* ULN vs > 3*ULN). ‘Japanese’ is defined as race = Asian and country = Japan, ‘Chinese’ is defined as race = Asian and country = mainland China.”

~~Discussion During Meeting: FDA accepts Bayer’s proposal and stated that additional information requests, if necessary, will be submitted during the sNDA review.~~

~~Reference ID: 3964863 IND 075642 Page 10~~

~~b. Distribution of drug exposure(s) for the full population used in the analysis.~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

~~c. A summary table of final model parameters with their corresponding units.~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

~~d. Any plots deemed appropriate to support the clinical interpretation of modeling results.~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

~~e. A summary describing the clinical application of modeling results.~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

~~10. Conduct graphical analysis of time to first dose modification in the overall population and by exposure groups.~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

11. Please refer to the following guidance for general expectations on submitting pharmacometrics data and models: http://www.fda.gov/AboutFDA/CentersOffices/OfficeofMedicalProductsandTobacco/CDE R/ucm180482.htm.

~~Reference ID: 3964863 IND 075642 Page 11~~

~~Bayer’s Response (Received via email on July 20, 2016): Bayer accepts FDA’s responses and there is no further discussion.~~

~~Discussion During Meeting: No discussion occurred.~~

~~ADDITIONAL MEETING DISCUSSION~~

~~Labeling~~

Bayer inquired whether the safety data described in Section 5 of labeling could be consolidated across indications, rather than describing incidence rates separately for each indication. FDA stated that such an approach may be appropriate as long as the data pertaining to the specific Warnings are consistent in the different indications. FDA agreed that it could be appropriate to include the East Asian data in the pool if the data pertaining to the specific Warning are consistent with the data in the other populations, and providing the data are submitted in the sNDA.

~~PREA REQUIREMENTS~~

Under the Pediatric Research Equity Act (PREA) (21 U.S.C. 355c), all applications for new active ingredients (which includes new salts and new fixed combinations), new indications, new dosage forms, new dosing regimens, or new routes of administration are required to contain an assessment of the safety and effectiveness of the product for the claimed indication(s) in pediatric patients unless this requirement is waived, deferred, or inapplicable.

Because this drug product for this indication has an orphan drug designation, you are exempt from these requirements. Please include a statement that confirms this finding, along with a reference to this communication, as part of the pediatric section (1.9 for eCTD submissions) of your application. If there are any changes to your development plans that would cause your application to trigger PREA, your exempt status would change.

~~DATA STANDARDS FOR STUDIES~~

Under section 745A(a) of the FD&C Act, electronic submissions “shall be submitted in such electronic format as specified by [FDA].” FDA has determined that study data contained in electronic submissions (i.e., NDAs, BLAs, ANDAs and INDs) must be in a format that the Agency can process, review, and archive. Currently, the Agency can process, review, and archive electronic submissions of clinical and nonclinical study data that use the standards specified in the Data Standards Catalog (Catalog) (See http://www.fda.gov/forindustry/datastandards/studydatastandards/default.htm).

On December 17, 2014, FDA issued final guidance, Providing Electronic Submissions in Electronic Format--- Standardized Study Data (http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ UCM292334.pdf). This guidance describes the submission types, the standardized study data

~~Reference ID: 3964863 IND 075642 Page 12~~

requirements, and when standardized study data will be required. Further, it describes the availability of implementation support in the form of a technical specifications document, Study Data Technical Conformance Guide (Conformance Guide) (See http://www.fda.gov/downloads/ForIndustry/DataStandards/StudyDataStandards/UCM384744.pd f), as well as email access to the eData Team ([email protected]) for specific questions related to study data standards. Standardized study data will be required in marketing application submissions for clinical and nonclinical studies that start on or after December 17, 2016. Standardized study data will be required in commercial IND application submissions for clinical and nonclinical studies that start on or after December 17, 2017. CDER has produced a Study Data Standards Resources web page that provides specifications for sponsors regarding implementation and submission of clinical and nonclinical study data in a standardized format. This web page will be updated regularly to reflect CDER's growing experience in order to meet the needs of its reviewers.

Although the submission of study data in conformance to the standards listed in the FDA Data Standards Catalog will not be required in studies that start before December 17, 2016, CDER strongly encourages IND sponsors to use the FDA supported data standards for the submission of IND applications and marketing applications. The implementation of data standards should occur as early as possible in the product development lifecycle, so that data standards are accounted for in the design, conduct, and analysis of clinical and nonclinical studies. For clinical and nonclinical studies, IND sponsors should include a plan (e.g., in the IND) describing the submission of standardized study data to FDA. This study data standardization plan (see the Conformance Guide) will assist FDA in identifying potential data standardization issues early in the development program.

~~Additional information can be found at http://www.fda.gov/Drugs/DevelopmentApprovalProcess/FormsSubmissionRequirements/Electr onicSubmissions/ucm248635.htm.~~

For general toxicology, supporting nonclinical toxicokinetic, and carcinogenicity studies, CDER encourages sponsors to use Standards for the Exchange of Nonclinical Data (SEND) and submit sample or test data sets before implementation becomes required. CDER will provide feedback to sponsors on the suitability of these test data sets. Information about submitting a test submission can be found here: http://www.fda.gov/Drugs/DevelopmentApprovalProcess/FormsSubmissionRequirements/Electr onicSubmissions/ucm174459.htm.

LABORATORY TEST UNITS FOR CLINICAL TRIALS CDER strongly encourages IND sponsors to identify the laboratory test units that will be reported in clinical trials that support applications for investigational new drugs and product registration. Although Système International (SI) units may be the standard reporting mechanism globally, dual reporting of a reasonable subset of laboratory tests in U.S. conventional units and SI units might be necessary to minimize conversion needs during review. Identification of units to be used for laboratory tests in clinical trials and solicitation of input from the review divisions should occur as early as possible in the development process. For more information, please see the FDA website entitled, Study Data Standards Resources and the

~~Reference ID: 3964863 IND 075642 Page 13~~

~~CDER/CBER Position on Use of SI Units for Lab Tests website found at http://www.fda.gov/ForIndustry/DataStandards/StudyDataStandards/ucm372553.htm.~~

~~ABUSE POTENTIAL ASSESSMENT~~

Drugs that affect the central nervous system, are chemically or pharmacologically similar to other drugs with known abuse potential, or produce psychoactive effects such as mood or cognitive changes (e.g., euphoria, hallucinations) need to be evaluated for their abuse potential and a proposal for scheduling will be required at the time of the NDA submission [21 CFR 314.50(d)(5)(vii)]. For information on the abuse potential evaluation and information required at the time of your NDA submission, see the draft guidance for industry, Guidance for Industry Assessment of Abuse Potential of Drugs, available at: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/U CM198650.pdf.

~~Office of Scientific Investigations (OSI) Requests~~

The Office of Scientific Investigations (OSI) requests that the following items be provided to facilitate development of clinical investigator and sponsor/monitor/CRO inspection assignments, and the background packages that are sent with those assignments to the FDA field investigators who conduct those inspections (Item I and II). This information is requested for all major trials used to support safety and efficacy in the application (i.e., phase 2/3 pivotal trials). Please note that if the requested items are provided elsewhere in submission in the format described, the Applicant can describe location or provide a link to the requested information.

The dataset that is requested in Item III below is for use in a clinical site selection model that is being piloted in CDER. Electronic submission of the site level dataset is voluntary and is intended to facilitate the timely selection of appropriate clinical sites for FDA inspection as part of the application and/or supplement review process. This request also provides instructions for where OSI requested items should be placed within an eCTD submission (Attachment 1, Technical Instructions: Submitting Bioresearch Monitoring (BIMO) Clinical Data in eCTD Format).

I. Request for general study related information and comprehensive clinical investigator information (if items are provided elsewhere in submission, describe location or provide link to requested information).

1. Please include the following information in a tabular format in the original NDA for each of the completed pivotal clinical trials: a. Site number b. Principal investigator c. Site Location: Address (e.g., Street, City, State, Country) and contact information (i.e., phone, fax, email) d. Location of Principal Investigator: Address (e.g., Street, City, State, and Country) and contact information (i.e., phone, fax, email). If the Applicant is aware of changes to a

~~Reference ID: 3964863 IND 075642 Page 14~~

~~clinical investigator’s site address or contact information since the time of the clinical investigator’s participation in the study, we request that this updated information also be provided.~~

2. Please include the following information in a tabular format, by site, in the original NDA for each of the completed pivotal clinical trials: a. Number of subjects screened at each site b. Number of subjects randomized at each site c. Number of subjects treated who prematurely discontinued for each site by site

3. Please include the following information in a tabular format in the NDA for each of the completed pivotal clinical trials: a. Location at which sponsor trial documentation is maintained (e.g., , monitoring plans and reports, training records, data management plans, drug accountability records, IND safety reports, or other sponsor records as described ICH E6, Section 8). This is the actual physical site(s) where documents are maintained and would be available for inspection b. Name, address and contact information of all Contract Research Organization (CROs) used in the conduct of the clinical trials and brief statement of trial related functions transferred to them. If this information has been submitted in eCTD format previously (e.g., as an addendum to a Form FDA 1571, you may identify the location(s) and/or provide link(s) to information previously provided. c. The location at which trial documentation and records generated by the CROs with respect to their roles and responsibilities in conduct of respective studies is maintained. As above, this is the actual physical site where documents would be available for inspection.

4. For each pivotal trial, provide a sample annotated Case Report Form (or identify the location and/or provide a link if provided elsewhere in the submission). 5. For each pivotal trial provide original protocol and all amendments ((or identify the location and/or provide a link if provided elsewhere in the submission).

~~II. Request for Subject Level Data Listings by Site~~

1. For each pivotal trial: Site-specific individual subject data listings (hereafter referred to as “line listings”). For each site, provide line listings for: a. Listing for each subject consented/enrolled; for subjects who were not randomized to treatment and/or treated with study therapy, include reason not randomized and/or treated b. Subject listing for treatment assignment (randomization) c. Listing of subjects that discontinued from study treatment and subjects that discontinued from the study completely (i.e., withdrew consent) with date and reason discontinued d. Listing of per protocol subjects/ non-per protocol subjects and reason not per protocol e. By subject listing of eligibility determination (i.e., inclusion and exclusion criteria)

~~Reference ID: 3964863 IND 075642 Page 15~~

f. By subject listing, of AEs, SAEs, deaths and dates g. By subject listing of protocol violations and/or deviations reported in the NDA, including a description of the deviation/violation h. By subject listing of the primary and secondary endpoint efficacy parameters or events. For derived or calculated endpoints, provide the raw data listings used to generate the derived/calculated endpoint. i. By subject listing of concomitant medications (as appropriate to the pivotal clinical trials) j. By subject listing, of testing (e.g., laboratory, ECG) performed for safety monitoring

~~2. We request that one PDF file be created for each pivotal Phase 2 and Phase 3 study using the following format:~~

~~III. Request for Site Level Dataset:~~

OSI is piloting a risk based model for site selection. Voluntary electronic submission of site level datasets is intended to facilitate the timely selection of appropriate clinical sites for FDA inspection as part of the application and/or supplement review process. If you wish to voluntarily provide a dataset, please refer to the draft Guidance for Industry Providing Submissions in Electronic Format – Summary Level Clinical Site Data for CDER’s Inspection Planning” (available at the following link http://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/FormsSubmissionRequire ments/UCM332468.pdf ) for the structure and format of this data set.

~~Reference ID: 3964863 IND 075642 Page 16~~

~~Attachment 1 Technical Instructions: Submitting Bioresearch Monitoring (BIMO) Clinical Data in eCTD Format~~

A. Data submitted for OSI review belongs in Module 5 of the eCTD. For items I and II in the chart below, the files should be linked into the Study Tagging File (STF) for each study. Leaf titles for this data should be named “BIMO [list study ID, followed by brief description of file being submitted].” In addition, a BIMO STF should be constructed and placed in Module 5.3.5.4, Other Study reports and related information. The study ID for this STF should be “bimo.” Files for items I, II and III below should be linked into this BIMO STF, using file tags indicated below. The item III site-level dataset filename should be “clinsite.xpt.”

DSI Pre- STF File Tag Used For Allowable NDA File Request Formats Item1 I data-listing-dataset Data listings, by study .pdf I annotated-crf Sample annotated case .pdf report form, by study II data-listing-dataset Data listings, by study .pdf (Line listings, by site) III data-listing-dataset Site-level datasets, across .xpt studies III data-listing-data-definition Define file .pdf

~~B. In addition, within the directory structure, the item III site-level dataset should be placed in the M5 folder as follows:~~

C. It is recommended, but not required, that a Reviewer’s Guide in PDF format be included. If this Guide is included, it should be included in the BIMO STF. The leaf title should be “BIMO Reviewer Guide.” The guide should contain a description of the BIMO elements being submitted with hyperlinks to those elements in Module 5.

~~1 Please see the OSI Pre-NDA/BLA Request document for a full description of requested data files~~

~~Reference ID: 3964863 IND 075642 Page 17~~

~~References:~~

~~eCTD Backbone Specification for Study Tagging Files v. 2.6.1 (http://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/FormsSubmissionRequire ments/ElectronicSubmissions/UCM163560.pdf)~~

~~FDA eCTD web page (http://www.fda.gov/Drugs/DevelopmentApprovalProcess/FormsSubmissionRequirements/Elect ronicSubmissions/ucm153574.htm)~~

~~For general help with eCTD submissions: [email protected]~~

NEW PROTOCOLS AND CHANGES TO PROTOCOLS To ensure that the Division is aware of your continued drug development plans and to facilitate successful interactions with the Division, including provision of advice and timely responses to your questions, we request that the cover letter for all new phase 2 or phase 3 protocol submissions to your IND or changes to these protocols include the following information:

1. Study phase 2. Statement of whether the study is intended to support marketing and/or labeling changes 3. Study objectives (e.g., dose finding) 4. Population 5. A brief description of the study design (e.g., placebo or active controlled) 6. Specific concerns for which you anticipate the Division will have comments 7. For changes to protocols only, also include the following information: • A brief summary of the substantive change(s) to the protocol (e.g., changes to endpoint measures, dose, and/or population) • Other significant changes • Proposed implementation date

~~We recommend you consider requesting a meeting to facilitate discussion of multiple and/or complex issues.~~

~~ISSUES REQUIRING FURTHER DISCUSSION No issues requiring further discussion.~~

~~ACTION ITEMS No action items.~~

~~ATTACHMENTS AND HANDOUTS “Stivarga (regorafenib) Study 15982 (RESORCE) Key Results”~~

~~14 Page(s) has been Withheld in Full as b4 (CCI/TS) immediately following this page~~

Reference ID: 3964863 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------IDARA UDOH 07/27/2016

~~Reference ID: 3964863~~